TMCnet News
CallWall: Tracking Resident Calls to Improve Clinical Utilization of Pathology Laboratories [Archives of Pathology & Laboratory Medicine](Archives of Pathology & Laboratory Medicine Via Acquire Media NewsEdge) Context.-Clinical pathology (CP) laboratories are used for millions of tests each year. These lead to thousands of calls to CP residents. However, although laboratory utilization is a frequent topic of study, clinical utilization-the content of the interactions between clinicians and CP residents-is not. Because it reflects questions about laboratory utilization, clinical utilization could suggest ways to improve both training and care by reducing diagnostic error. Objectives.-To build and implement a secure, scalable Web-based system to allow CP residents at any hospital to track the calls they receive, the interaction's context, and the action taken as a result, with evidence where applicable, and to use this system to report on clinical utilization at a major academic hospital. Design.-Entries were analyzed from a nearly year-long period to describe the clinical utilization of CP at a large academic teaching hospital. Results.-Sixteen residents logged 847 calls during 10 months, roughly evenly distributed among transfusion medicine, chemistry, microbiology, and hematopathology. Calls covered 94 different analytes in chemistry and 71 different organisms or tests in microbiology. Analysis revealed areas where CP can improve clinical care through educating the clinical services, for example, about ordering Rh immune globulin, testosterone testing, and diagnosis of tick-borne diseases. Documenting calls also highlighted patterns among residents. Conclusions.-Clinical utilization is a potentially rich knowledge base for improving patient care and resident training. Our resident call-tracking system is a useful way for measuring clinical utilization and mining it for actionable information. (Arch Pathol Lab Med. 2011;135:920-924) The services that the clinical pathology (CP) laboratories provide touch all medical specialties. This central position has led to interest in laboratory utilization-how and when clinicians order tests-from many specialties1,2 and from within pathology itself.3 Less well studied is how nonpathologists interact with clinical pathologists on clinical issues, which we refer to here as clinical utilization. The classic example of clinical utilization is when a nonpathology specialist-for example, a medicine resident or attending-pages or calls the CP resident-on-call about how to interpret a test result or whether to order a test or procedure. As a rule, the clinicians who place these calls are already good at interpreting tests and routinely consult a range of print and electronic resources before involving others. Hence when they page CP residents, it tends to be only for cases that are complicated or for which ready guidance is hard to find. These are cases that are likely to lead to diagnostic error. From CP's perspective, handling calls is an essential part of resident training.4 As a result, CP resident calls have potential as a unique resource for understanding clinical utilization as well as for CP teaching.5 Realizing this potential requires that calls be documented electronically in central, searchable databases. It also requires that calls be easy for residents to enter. At least one CP training program has developed such a database that is accessible over the Web, making calls easy to enter from anywhere.6 This system uses a Microsoft Access (Microsoft, Redmond, Washington) back end, which is suitable for local prototyping but not for enterprise-class distribution because of its size and stability limitations. Nevertheless, it represents a marked improvement over most training programs: At present 76% lack electronic call-tracking systems entirely.7 Several programs have used their systems to report a breakdown of calls by administrative category: mislabeled specimens, requisition clarifications, and so forth.6,8 But to our knowledge, none have reported on clinical utilization, that is, the clinical content of those calls. Here we report on clinical utilization in CP during 10 months at a large academic hospital and describe CallWall, the enterprise-scalable, best-practices, Webbased system that we developed for tracking calls. Information on implementing the system at other institutions is updated regularly at http://www.call-wall.org/ about (Accessed September 3, 2010). MATERIALS AND METHODS Hospital Setting The Beth Israel-Deaconess Medical Center (Boston, Massachusetts) is a 621-bed tertiary care center and a teaching hospital of Harvard Medical School. In our department, approximately 10 CP residents rotate through the 4 core laboratories of transfusion medicine (blood bank), chemistry, microbiology, and hematopathology, as well as molecular pathology and cytogenetics, with 4 to 6 going through core rotations and the remainder joining for advanced rotations from time to time. They were instructed at weekly conference to begin using the new system to track their calls, that attending pathologists would be checking the system to ensure compliance, and that the call list for conference would now come from the system, and were reminded of this individually and collectively multiple times during the course of the study period. Design We prototyped our system using a Microsoft Access database that resided on a hospital-based server. The database was accessible from throughout the hospital on machines running Microsoft Windows XP or from outside the hospital through virtual private network connection. Fields in the call-entry form were maintained essentially unchanged in the transition to the Web-based system (Figure 1), with the home page presenting a list of calls in report format and including controls for customizing the report by laboratory and date range (Figure 2). For scalability, and particularly so that other institutions could easily implement the system for their programs, we transitioned to MySQL (Oracle Corporation, Redwood Shores, California) for our Web-based system. For flexibility, we built this system using the open-source Web application framework Ruby on Rails. For stability, we developed it using version control on a multiply redundant server system (ProjectLounge, Somerville, Massachusetts; www.projectlounge.com. Accessed September 18, 2010) hosted in a professionally maintained colocation center. For security, each user was given his or her own username-password combination. And for patient safety, medical record numbers were scrambled for display purposes, residents were instructed not to enter patient names or other protected health information, and the data in this teaching database were kept separate from the patient record. These represent key innovations over previous systems. A department faculty member served as system administrator. RESULTS Sixteen residents logged 847 calls during the 10-month period from September 2009 through June 2010 (Table 1), for an average of 85 calls per month (range, 64-112). These were roughly evenly distributed among transfusion medicine (271 calls; 32% of total), chemistry (213 calls; 25%), microbiology (205 calls; 24%), and hematology (158 calls; 19%). The 4 residents on full-time CP service accounted for most calls (613 calls; 72%). Calls from laboratorians often led to conversations with the ordering clinician. Transfusion medicine calls ran the gamut of specialty expertise (Table 2). Eight areas accounted for threefourths of calls, with requests for apheresis procedures being the single most common issue (40 calls; 15% of all transfusion medicine calls), followed by questions about Rh immune globulin (29 calls; 11%) and platelets (28 calls; 10%). About 20% of calls came directly or indirectly from the blood bank for help or approval in managing bloodproduct inventory. An example of a call was to evaluate a patient with symptoms and biopsy evidence of thrombotic microangiopathy for thrombotic thrombocytopenic purpura, for which apheresis would be useful, versus malignant hypertension, for which it would not. The resident recorded that the patient was treated with apheresis pending results on ADAMTS13, in accordance with 3 articles found on this issue; the resident also recorded the citations for these articles in the system.9-11 Calls in chemistry covered 94 different analytes (Table 2). The most common area of inquiry was endocrine testing with 37 calls (17% of all chemistry calls); most of these were regarding testosterone (11 calls) or thyroid conditions (10 calls). Next was gastroenterology with 35 calls (16%), 15 of which were about testing for hepatitis viruses. Urgent care issues were third with 20 calls (9%), mostly regarding drug screening (alcohol, opiates, amphetamines, etc). Human immunodeficiency virus and antibiotic-level testing accounted for 17 calls (8%), as did calls about renal issues. The remaining 50% of calls spanned diverse areas, such as pregnancy and fetal maturity (7 calls; 3%). The 205 microbiology calls covered 71 different organisms or tests. A fourth of calls were about just 5 organisms: human immunodeficiency virus (13 calls; 6%), Borrelia (12 calls; 6%), cytomegalovirus (10 calls; 5%), human herpesvirus 6 (9 calls; 4%), and hepatitis D virus (8 calls; 4%). Only 9 calls (4%) were about drug susceptibility testing; an additional 9 calls were for consults on anatomic pathology specimens. One in every 8 calls (26 calls; 13%) was about testing for tick-borne diseases (Anaplasma, Babesia, Borrelia, Ehrlichia), split evenly between the warmer and cooler months. Almost half of all calls were requests for sendout testing (96 calls; 47%), the highest of any of the core laboratories. About a third (28 calls) of these were either declined or else modified after the resident spoke with the ordering clinician; the reasons for and outcomes of these conversations were recorded in the system for future reference. Finally, almost half (74 calls; 47%) of the 158 recorded calls in hematopathology were for review of atypical cells flagged by technologists. Another 18 calls were about factor V Leiden (10 calls; 6%) or JAK2 (8 calls; 5%) testing. Seventeen calls (11%) were about specimen handling; most of these were requests from clinicians to rush samples, but they also included unusual cases, such as detection of sperm in the urine of a young girl, and clinicians seeking education, such as how to square a negative blood urine dipstick with report of 16 red blood cells on microscopy. Nine calls (6%) regarded critical values, all but 3 of which were for new findings of leukemic cells. In contrast to the other 3 core laboratories, in hematopathology different residents appeared to log different kinds of calls: Of the 21 hematopathology entries that recorded interesting teaching points, all but one were from 1 resident; in contrast, 42 of 57 calls (74%) from another resident were for review of atypical cells. Three calls not included in the hematopathology total were clearly mislabeled (2 were microbiology, 1 chemistry), and another 3 (2% of hematopathology calls; ,1% of all calls) were too cryptic to be interpreted on review. COMMENT Calls placed to CP residents are important for training; the clinical utilization they describe is also potentially valuable for improving clinical care and specifically for addressing diagnostic error across medicine.12-16 Two broad requirements are necessary to realize this potential. First, calls must be easy for residents to log and search. Our online system is as easy as Webmail: a simple username-password combination to log in, a single button to create a call, a simple but comprehensive search field; no special installations are necessary, and it works on any machine with a Web browser and Internet connection. It is similarly robust, built according to enterprise-software development standards. Second, residents must want to track their calls: They must find it helpful and/or feel pressure to track calls from peers or faculty. Search capability helps residents find answers to questions pertinent to their calls (have there been cases like this before, what was done then, and why). Also helping spur adoption was the expectation by faculty that the call tracker would be the source of calls for weekly conference. In developing CallWall, we faced a choice between building de novo and customizing an existing issuetracking or call-center service system. Many such systems exist17; their number speaks to how each area of need (or ''vertical market,'' in business terminology) demands somewhat different features. Even some standard features- such as listing start and end times, tracking customer e-mails, and displaying company information- do not readily apply to resident call tracking. To avoid distracting the CallWall user, in a customized system, these would have to be hidden or removed. Based on the experience of one of us (I.M.C.) building such systems at IBM (Armonk, New York), we concluded that the effort required to customize an existing solution in this way would exceed the work required to develop a system de novo. In this case, de novo development would achieve a more tailored solution in less time, with fewer lines of code and thus fewer bugs, and so we took this approach. Overall, implementation was deemed successful. Personal experience suggests the calls entered are a substantial fraction, if still an underestimate, of all calls received. Heterogeneity among residents on the hematopathology service in the kind of calls entered suggests better guidance can be given to residents as to what calls to track. Although preliminary evidence suggests that use of CallWall has led to improved quality control procedures in the department, the present study is limited to describing the system itself and the nature of this first set of calls. That said, the corpus of reported calls comprises a substantial database that reflects the clinical utilization of CP, and thus the needs of the hospital, during the study period and revealed trends that were not apparent during day-to-day operations. Examples include the frequency of calls relating to testosterone and thyroid conditions on the chemistry service and of tick-borne infections on the microbiology service. These suggest it might be helpful to educate clinicians on the medical services regarding utilization for these specific issues, and that this might lead to fewer errors in these areas. Review of our calls also suggests ways to improve data collection. For example, the breakdown of calls suggests service-specific text-entry fields, such as type-ahead fields for organism or test name for microbiology and chemistry, would help speed call entry, search, and analysis; adding support for images would also be useful. These improvements are underway. As call tracking becomes more pervasive and standardized, we hope it will help formalize CP accreditation by providing a record of cases residents have seen. It can also help residents guide their own teaching, by letting them see at a glance the topics they have encountered often and request more exposure to those seen rarely. By adhering to software design best practices, we have made it straightforward to add functionality to improve the system and also to make it available-in a password-protected, internal review board-compliant fashion-to hospitals elsewhere. In this way, we hope the ability to easily document and analyze clinical utilization can help improve training and practice in CP. References 1. Fiappo E, Quattrin R, Calligaris L, Cappelletti P, Brusaferro S. Appropriate laboratory utilization in diagnosing pulmonary embolism. Ann Clin Biochem. 2009;46:18-23. 2. Akan P, Cimrin D, Ormen M, et al. The inappropriate use of HbA1c testing to monitor glycemia: is there evidence in laboratory data? J Eval Clin Pract. 2007; 13(1):21-24. 3. Emerson JF, Emerson SS. The impact of requisition design on laboratory utilization. Am J Clin Pathol. 2001;116(6):879-884. 4. Spitzer, ED, Pierce GF, McDonald JM. A laboratory medicine residency training program that includes clinical consultation and research. Arch Pathol Lab Med. 1990;114(4):360-362. 5. Luning Prak ET, Young DS, Kamoun M, et al. 2008 ACLPS panel discussion on resident education in clinical pathology. Am J Clin Pathol. 2009;131(5):618- 622. 6. Hoofnagle AN, Chou D, Astion ML. Online database for documenting clinical pathology resident education. Clin Chem. 2007;53(1):134-137. 7. Gronowski AM. Innovations in ComACC-approved clinical chemistry training programs. Paper presented at: 45th Annual ACLPS Meeting; June 3, 2010; Nashville, TN. 8. Do MC, Ben-Ezra J, McPherson RA. Call subject patterns among on-call clinical pathology residents in an academic institution: how can tracking changes in patterns over time benefit resident education? Arch Pathol Lab Med. 2008; 132(8):1317-1320. 9. Shibagaki Y, Fujita T. Thrombotic microangiopathy in malignant hypertension and hemolytic uremic syndrome (HUS)/thrombotic thrombocytopenic purpura (TTP): can we differentiate one from the other? Hypertens Res. 2005; 28(1):89-95. 10. Egan JA, Bandarenko N, Hay SN, et al. Differentiating thrombotic microangiopathies induced by severe hypertension from anemia and thrombocytopenia seen in thrombotic thrombocytopenia purpura. J Clin Apheresis. 2004; 19(3):125-129. 11. Zhang B, Xing C, Yu X, Sun B, Zhao X, Qian J. Renal thrombotic microangiopathies induced by severe hypertension. Hypertens Res. 2008;31(3):479-483. 12. Leape LL. Errors in medicine. Clin Chim Acta. 2009;404(1):2-5. 13. Gandhi TK, Kachalia A, Thomas EJ, et al. Missed and delayed diagnoses in the ambulatory setting: a study of closed malpractice claims. Ann Intern Med. 2006;145(7):488-496. 14. Leape LL, Brennan TA, Laird N, et al. The nature of adverse events in hospitalized patients. Results of the Harvard Medical Practice Study II. N Engl J Med. 1991;324(6):377-384. 15. Schiff GD, Hasan O, Kim S, et al. Diagnostic error in medicine: analysis of 583 physician-reported errors. Arch Intern Med. 2009;169(20):1881- 1887. 16. Newman-Toker DE, Pronovost PJ. Diagnostic errors-the next frontier for patient safety. J Am Med Assoc. 2009;301(10):1060-1062. 17. Wikipedia. Available at: http://en.wikipedia.org/wiki/Comparison_of_ issue-tracking_systems. Accessed September 3, 2010. Thomas P. Buck, MD; Ian M. Connor, B App Sci, LLB; Gary L. Horowitz, MD; Ramy A. Arnaout, MD, DPhil Accepted for publication September 22, 2010. From the Department of Pathology, Beth Israel-Deaconess Medical Center, Boston, Massachusetts (Drs Buck, Horowitz, and Arnaout); and ProjectLounge.com Inc. Cambridge, Massachusetts (Mr Connor). Drs Buck, Horowitz, and Arnaout have no relevant financial interest in the products or companies described in this article. Reprints: Ramy A. Arnaout, MD, DPhil, Department of Pathology, Beth Israel-Deaconess Medical Center, 330 Brookline Ave, YA309, Boston, MA 02215 (e-mail: [email protected]). (c) 2011 College of American Pathologists |