Clinical Chemistry 56:3 417–423 (2010)
Informatics and Statistics
Implementation of a Closed-Loop Reporting System for Critical Values and Clinical Communication in Compliance with Goals of The Joint Commission Fritz F. Parl,1* Mandy F. O’Leary,1 Allen B. Kaiser,2 John M. Paulett,3 Kristina Statnikova,3 and Edward K. Shultz1,3
BACKGROUND: Current practices of reporting critical laboratory values make it challenging to measure and assess the timeliness of receipt by the treating physician as required by The Joint Commission’s 2008 National Patient Safety Goals. METHODS:
A multidisciplinary team of laboratorians, clinicians, and information technology experts developed an electronic ALERTS system that reports critical values via the laboratory and hospital information systems to alphanumeric pagers of clinicians and ensures failsafe notification, instant documentation, automatic tracking, escalation, and reporting of critical value alerts. A method for automated acknowledgment of message receipt was incorporated into the system design.
RESULTS:
The ALERTS system has been applied to inpatients and eliminated approximately 9000 phone calls a year made by medical technologists. Although a small number of phone calls were still made as a result of pages not acknowledged by clinicians within 10 min, they were made by telephone operators, who either contacted the same physician who was initially paged by the automated system or identified and contacted alternate physicians or the patient’s nurse. Overall, documentation of physician acknowledgment of receipt in the electronic medical record increased to 95% of critical values over 9 months, while the median time decreased to ⬍3 min.
CONCLUSIONS: We improved laboratory efficiency and physician communication by developing an electronic system for reporting of critical values that is in compliance with The Joint Commission’s goals.
© 2009 American Association for Clinical Chemistry
Departments of 1 Pathology and 2 Internal Medicine and 3 Informatics Center, Vanderbilt University Medical Center, Nashville, TN. * Address correspondence to this author at: Clinical Chemistry Laboratory, Professor of Pathology, Department of Pathology, TVC 4918, Vanderbilt University Medical Center, Nashville, TN 37232. Fax 615-343-9563; e-mail fritz.
[email protected].
Ever since George Lundberg introduced the concept of critical laboratory values in the 1970s (1 ), medical technologists have considered it their duty to expeditiously report these values by calling the respective patient units and reading the numbers to the caretaker. Critical value reporting became official when it was incorporated into the CLIA of 1988, and the CLIA recommendations were widely adopted by laboratoryaccrediting agencies, such as The Joint Commission and the College of American Pathologists (CAP)4 (2, 3 ). Subsequently, the requirement of having the caregiver write down and read back the critical value information (so-called read-back) was introduced as a means of improving patient safety (4, 5 ). Clearly the reporting of critical results can be improved. CAP conducted a study across 623 institutions and found that slightly ⬎5% of critical alerts were “abandoned,” i.e., not reported to anyone who would accept responsibility for the critical value (6 ). In a follow-up study of 121 institutions, the median notification time for inpatients was 4 min (7 ). However, this time was needed to notify a nonlicensed caregiver; an additional 3 min were required to notify a physician, adding up to a total of 7 min. The amount of time spent by laboratories reporting thousands of crisis values per year is significant in hospitals that treat many critically ill patients. In recognition of the substantial commitment of effort and time, a few attempts have been made to develop automated systems for the reporting of critical values (8, 9 ). In reporting critical values for inpatients, medical technologists generally call the floor where the patient is located. Typically a clerk will answer the phone and the technologist will then request to speak to a nurse or physician familiar with the patient. The recipient will write down the result and read it back to the technologist. In the case where the recipient is a nurse, the nurse
Received August 12, 2009; accepted October 28, 2009. Previously published online at DOI: 10.1373/clinchem.2009.135376 4 Nonstandard abbreviations: CAP, College of American Pathologists; VUMC, Vanderbilt University Medical Center; EMR, electronic medical record; CPOE, computer physician order entry; FYI, for your information; LIS, Laboratory Information System.
417
will report the critical result to the patient’s physician with read-back and record the reporting in the nurses’ notes or other component of the medical record. Audits have shown that receipt of the critical value by the clinician is documented in only about two thirds of cases. It is unknown how much of this represents an abandoned effort to contact a physician or simply a lack of documentation (10 ). The purpose of reporting a crisis value is to alert the caregiver, typically a physician, who will be able to act on the information. Thus, documentation of the receipt of the critical values by the caregiver is as important as assurance of rapid delivery. Accordingly, The Joint Commission now requires health care organizations to track and improve the timeliness of reporting and receipt of critical test results by the responsible, licensed clinician. Specifically Goal 2 of The Joint Commission’s 2007 and 2008 National Patient Safety Goals states that laboratories must measure and assess both the timeliness of delivery and receipt of critical values by the licensed physician. At the present time, many hospitals struggle with creating the receipt, as it often requires a nurse to stop and document the details of a physician contact. Monitoring of the physician receipt and its timing is typically done by a noncomprehensive manual chart audit. We developed an electronic ALERTS system that reports critical values via the laboratory and hospital information systems to alphanumeric pagers of clinicians. The system improved laboratory efficiency and clinical communications and thereby addressed regulatory, medico-legal, practical workflow, and clinical concerns. Materials and Methods SETTING
The ALERTS system was developed at Vanderbilt University Medical Center (VUMC), a tertiary-care 1000bed academic medical center in Nashville, TN, as a collaborative effort between the Clinical Diagnostic Laboratories in the Department of Pathology, the Informatics Center, and the main clinical departments (Internal Medicine, Pediatrics, Surgery, and Emergency Medicine). Another component critical to the successful adoption of the system was the participation of the Telephone Operator Services in the Information Technology Services Department. CLINICAL DIAGNOSTIC LABORATORIES
The Core Laboratory, which combines clinical chemistry, coagulation, and hematology, uses a list of critical values that are similar to values reported in the literature (11 ). The list of critical values flagged in the laboratory information system is identical to that used by 418 Clinical Chemistry 56:3 (2010)
the medical technologists for calling the patient units and used by the ALERTS system (Table 1). INFORMATICS CENTER
The Informatics Center supports the clinical information needs of all VUMC inpatient and outpatient services. The ALERTS system’s architectural design and software and hardware development was done within the Informatics Center. CLINICAL DEPARTMENTS
More than 1200 physicians practice in the Vanderbilt Medical Group at VUMC, supervising the training of about 850 residents and fellows. Residents and fellows order most of the laboratory tests. The ALERTS system was applied only to samples originating from inpatients. TELEPHONE OPERATOR SERVICES
The operators maintain a database of digital pagers for approximately 5000 medical center and university personnel, provide pager forwarding services for admitting, consult, and team pagers, maintain and provide daily on-call information for more than 200 medical center clinical services to physicians, nursing staff, and admitting and emergency services. Because operators have up-to-date on-call schedules for all clinical services, they are in a unique position to identify alternate physicians who need to be informed about a critical value in case the primary doctor does not acknowledge alert receipt. A cadre of 14 of the most experienced operators was selected and trained in using the ALERTS system to monitor critical value alerts acknowledgment, to close escalated pages, and to be proficient in reading and pronouncing the names of the critical laboratory tests. Operators’ proficiency in communicating critical value information is periodically audited by a representative from the Clinical Diagnostic Laboratories. Results The multidisciplinary team designed a flowchart for the reporting of critical values and their acknowledgment (Fig. 1). The flowchart starts with the medical technologist entering a test result into the laboratory information system (GE Healthcare Triple G). The ALERTS system captures the outgoing HL-7 result and, if the result fulfills the critical value criteria defined in Table 1, sends a critical value alert via the text pager system. The pager number of a covering physician for the patient has been assigned during the creation of admitting orders or the patient transfer process. The text message lists the result with reference
Closed-Loop Reporting System for Critical Values
Table 1. Critical laboratory values. Test
Unit (SI)
Low limit
High limit
Clinical chemistry (adult) CO2 (bicarbonate)
mmol/L
10
40
Calcium
mg/dL (mmol/L)
6.0 (1.5)
13.0 (3.25)
Calcium, ionized
mg/dL (mmol/L)
3.0 (0.75)
6.5 (1.63)
Glucose
mg/dL (mmol/L)
40 (2.2)
500 (27.8)
Magnesium
mg/dL (mmol/L)
1.0 (0.50)
5.0 (2.5)
Phosphorus
mg/dL (mmol/L)
1.0 (0.32)
10.0 (3.2)
Potassium (plasma)
mEq/L (mmol/L)
2.5 (2.5)
6.2 (6.2)
Sodium
mEq/L (mmol/L)
115 (115)
160 (160)
Glucose tolerance
mg/dL (mmol/L)
60 (3.3)
500 (27.8)
CSF glucosea
mg/dL (mmol/L)
20 (1.1)
CSF protein
mg/dL (g/L)
300 (3.0)
Free T4
ng/dL (pmol/L)
3.5 (45.0)
TSH
U/mL (mIU/L)
100 (100)
mol/L (mol/L)
150 (150)
Clinical chemistry (pediatric) Ammonia Total bilirubin (newborn ⬍30 days)
mg/dL (mol/L)
Glucose (newborn ⬍30 days)
mg/dL (mmol/L)
30 (1.7)
200 (11.1)
Glucose (1 month to 1 year)
mg/dL (mmol/L)
30 (1.7)
300 (16.7)
Potassium (newborn ⬍10 days)
mEq/L (mmol/L)
2.0 (2.0)
7.0 (7.0)
Sodium (newborn ⬍10 days)
mEq/L (mmol/L)
125 (125)
160 (160)
mg/dL (mol/L)
60 (0.6)
18.0 (306)
Coagulation Fibrinogen PT (prothrombin time)/INR PTT (activated partial thromboplastin time)
7.0 s
200
Hematology Hemoglobin
g/dL (g/L)
6.0 (3.7)
%
18.0
Platelet count
⫻103/mm3 (⫻109/L)
30 and no previous within 7 days (30)
1000 and no previous within 6 months (1000)
White blood cell count
⫻103/mm3 (⫻109/L)
1.0 and no previous within 7 days (1.0)
100 and no previous within 30 days (100)
PCV (packed cell volume, hematocrit)
Peripheral smear morphology a
% blasts
69 (⬍2 months) 54 (ⱖ2 months)
10 in children, 5 in adults
CSF, cerebrospinal fluid; T4 , thyroxine; TSH, thyroid-stimulating hormone; INR, International Normalized Ratio.
range in addition to patient name, medical record number, and specimen collection timestamp. In the most common scenario, the physician acknowledges the page by calling the ALERTS touchtone phone response system using the phone extension listed in the message and dialing the acknowledgment code provided in the message. The physician’s identity is confirmed as part of the touchtone phone interaction. If his identity is not known by the system, he can use a unique code to identify himself. Alternatively, the phy-
sician can acknowledge the alert receipt electronically in the hospital electronic medical record (EMR) or computer physician order entry (CPOE) systems. A different pathway is initiated if the clinician does not respond to the critical value alert within 10 min (Fig. 1). The page is automatically escalated to the telephone operator who monitors the status of every alert in real time via the ALERTS Status Web Application. The operator contacts the responsible physician or the patient’s nurse by phone and communicates the critical Clinical Chemistry 56:3 (2010) 419
Fig. 1. Flowchart of critical value reporting and acknowledgment.
value information, receives a read-back, and documents the alert acknowledgment in the ALERTS system. A third scenario arises when the paged clinician does not know the patient and “rejects” the page (Fig. 1). In this case, the page is also escalated to Telephone Operator Services. The operator identifies an alternate physician using the up-to-date schedules of all clinical services and relates the message in a phone call with read-back by the physician. In the small number of cases where the operator has to relate critical value information for escalated alerts to the patient’s nurse owing to the unavailability of the patient’s physician, the nurse is then responsible for contacting the physician and informing him or her about the critical value and documenting the physician notification in the electronic nursing charting system. In all 3 pathways, the process ends with the physician acknowledging the receipt of the critical value, allowing him or her to take action and avert a 420 Clinical Chemistry 56:3 (2010)
negative clinical outcome. In case a sample with a critical value also has an interfering substance, a monitor informs the technologist of this situation. The technologist handles these cases manually, using the same logic as before the implementation of the ALERTS system. Table 2 summarizes the ALERTS data recorded and calculated electronically for the most recent 9 months of system use. Because the response times are skewed, we list the mean values as well as the median and 90% percentiles. In the following, we concentrate on the median because it is the parameter most commonly reported in the literature. After implementation of the ALERTS system, the median clinician response occurred in about 1.5 min. During the first 3 months, however, the responsible physicians acknowledged the pages using the automated options in only 70% of cases. To increase the physician acknowledgment rate, a “7-min reminder” page was implemented. Each time
Closed-Loop Reporting System for Critical Values
Table 2. Results of ALERTS system. Month 2008–2009 ALERTS matrix
Clinician pages, n Pages acknowledged, n (%)
Oct
825
Nov
Dec
728
Jan
631
721
Feb
758
Mar
779
Apr
681
May
June
795
933
573 (70) 503 (69) 435 (69) 586 (81) 659 (87) 622 (80) 592 (87) 700 (88) 829 (89)
Clinician response time, min Mean
2.6
2.3
2.3
2.9
3.0
2.7
2.9
3.0
2.9
Median
1.5
1.5
1.4
1.6
1.6
1.4
1.6
1.6
1.5
90% percentile
5.2
4.5
4.7
7.9
8.2
7.8
8.2
8.1
8.0
Pages escalated to operator, n (%)
252 (31) 225 (31) 196 (31) 135 (19)
99 (13) 157 (20)
89 (13)
95 (12) 104 (11)
Operator time to closure, min Mean
6.5
5.1
4.9
5.4
5.3
5.8
5.3
6.0
3.5
Median
5.3
4.0
3.9
3.8
3.9
4.3
3.4
4.0
3.0
10.6
9.6
9.1
10.6
90% percentile Escalated pages communicated to physician, n (%) Escalated pages communicated to nurse, n (%) Pages communicated to physician and documented by ALERTS, %
189 (23) 171 (24) 136 (22) 109 (15)
11.5
11.6
9.8
9.9
69 (9)
9.9
95 (12)
61 (9)
54 (7)
62 (7)
63 (8)
54 (7)
60 (10)
26 (4)
30 (4)
62 (8)
28 (4)
41 (5)
42 (5)
93
93
90
96
96
92
96
95
95
Time to ALERTS documentation of physician in EMR, min Mean
6.9
6.3
6.3
5.4
4.7
5.4
4.8
4.7
Median
2.9
2.3
2.4
2.2
2.0
2.0
1.8
2.0
1.7
17.4
15.4
16.2
14.2
13.0
14.5
13.2
12.6
11.8
90% percentile
a physician fails to acknowledge an alert within 7 min of being paged, the system sends a reminder page with a warning that failure to acknowledge the alert receipt in the next 3 min will result in the alert being escalated to the telephone operator. After the implementation of the 7-min reminder page, the physician acknowledgment rate increased to near 90%. The number of alerts escalated to telephone operators correspondingly decreased from 30% to 11% (Table 2). In 70% of these cases, the operators succeeded in closing escalated alerts by either contacting the original physician or identifying an alternate physician from the available on-call schedules. In a small fraction of cases where no physician was available (approximately 5% of the total alerts), the operators contacted the patient’s nurse to acknowledge the critical value page. The nurse was then responsible for contacting the physician and informing him or her about the critical value and documenting the physician notification in the electronic nursing charting system. The ALERTS system currently does not track documentation of these critical values, as this occurs between the nursing documentation system and the EMR. Excluding the escalation time of 10 min, the median response time achieved
4.1
by the operators was 5.3 min or less. Overall, documentation of physician acknowledgment of receipt in the electronic medical record increased to 95% of critical values over 9 months, while the median time decreased to ⬍3 min. Before implementing the ALERTS system, nurses were responsible for notifying the clinician about critical laboratory results, which they received via phone calls from the laboratory. After introducing the system, this notification duty was removed from the nursing personnel, alleviating their workload. Because nurses play a vital role in monitoring the patient condition, however, it was important to keep nursing personnel in the critical value information flow. Thus, a “for your information” (FYI) charge nurse paging functionality on the unit where the patient is located was added to the ALERTS system, sent at the same time a physician is alerted about a critical value. The FYI charge nurse pager message includes information about the critical value result and contact information of the physician notified by the system. The charge nurse FYI paging feature provides an additional patient safety mechanism by keeping nursing personnel in the information Clinical Chemistry 56:3 (2010) 421
flow, allowing nurses to intervene in patient care when immediate attention is required. The ALERTS system was implemented in stages, starting with a few tests and inpatient floors and ending with the entire hospital and all tests listed in Table 1. The full implementation of the ALERTS system has led to the near-complete elimination of phone calls by medical technologists reporting critical values for inpatients. Before the introduction of the ALERTS system, approximately 9000 phone calls per year were logged by the Clinical Laboratories. The time to contact the appropriate nurse on the patient floor varied from 1 to ⬎10 min. If we assume an average of 7 min per call (based on the study cited above (7 )) including readback of the result, the introduction of the ALERTS system resulted in time savings of 1050 h of technologists’ time for the Clinical Laboratories, or approximately one half of a full-time equivalent technologist. We did not measure any laboratory quality outcomes that might be affected by technologists having to call critical values; however, it is likely that the elimination of 9000 phone calls improved overall turnaround times and reduced the rate of clerical errors and/or corrected reports due to interruptions. Discussion A growing number of publications have addressed the reporting of critical values (4 –7, 9, 12–14 ). A recent CAP-sponsored study of 121 institutions determined that it takes a total of 7 min to relate a critical value to the responsible clinician (7 ). However, this determination, involving 3545 critical values, was performed manually in a labor-intensive manner. We report on 6850 critical values made over a 9-month period of time by the electronic closed-loop ALERTS system. The essential elements for a successful electronic reporting system have been defined (15 ). Shabot et al. (16 ) described a text pager– based notification of critical values, but the system did not document receipt of the alert. A conceptually similar system to ALERTS described by Kuperman et al. (8 ) also employed physician pagers and operators and nurses as part of an escalation strategy. The system was evaluated in a pilot study of 4 tests (glucose, potassium, sodium, hematocrit) monitored over a 2-month period. The documented physician response rate in this limited study was 69% (8 ). This is nearly identical to our direct physician response rate before we instituted the 7-min reminder, which resulted in an eventual 89% rate. Their system did not include a touchtone phone method for acknowledgment, which also may have contributed to our higher response rate, although the reminder appears more likely given the timing of the change. Im422 Clinical Chemistry 56:3 (2010)
portantly, their study differed from ours in that it did not report an overall physician receipt rate or response time. We were unable to identify any studies with documented physician receipt rates higher than 70%. We added the Telephone Operator Services as an integral component to increase documented direct physician contact to 95% (89% direct acknowledgment plus 6.5% via the operator). For the remaining 5%, the operators notified the nurse, allowing a 100% success rate of notification (no abandoned alerts). Although operators may not be an obvious choice, they turned out to be highly dedicated and well suited for this task. In contrast to laboratory-based medical technologists, operators have access to up-to-date call schedules for all clinical residents, fellows, and faculty and had minimal difficulty tracking a physician to redirect the text page or deliver the critical value by phone. Within a few months, adding the operators for escalation shortened the overall median time to physician documentation to ⬍3 min (Table 2), in contrast to the 7 min found in the CAP study (7 ). Although the operators had a noticeable impact on the success of the ALERTS system, it should be noted that the number of alerts escalated to them (about 100 per month or 3– 4 per day) was a negligible addition to their workload and had no adverse impact on operator performance. From the laboratory point of view, the ALERTS system has been a resounding success, with significant time saving by medical technologists. Similarly, the nursing staff has been relieved of the task of notifying the responsible clinician of critical values except in those few instances when the operator could not locate the physician. The physicians have accepted the ALERTS system and, after some initial hesitation and education, improved their acknowledgment rate. Instead of being verbally notified of a critical value by a nurse, the clinicians realize the benefit of the electronic alert and its built-in documentation. Finally, from an institutional point of view, the ALERTS system is entirely electronic and documents every time point in the reporting pathway. The ALERTS system has been successfully implemented for inpatients, providing information 24/7. Critical values for outpatients typically account for ⬍20% of all critical values (12, 14 ), and we plan to address these in the next phase of the project using the experience gained to date. Future work will also address the 5% of operator-to-nurse alerts. Currently the nurse is instructed to enter the physician notified within the nursing documentation system, which does not always occur. The plan is to automatically escalate the alert to the charge nurse if this documentation has not occurred in a reasonable amount of time. The ALERTS system captures the outgoing HL-7 results
Closed-Loop Reporting System for Critical Values
from our Laboratory Information System (LIS) but was developed as a freestanding program at VUMC without participation of the LIS vendor. Thus, the ALERTS system is amenable to use with any other LIS provided a standard HL-7 interface is accessible. The system is not commercially available at the present time. We conclude that the ALERTS system has decreased physician notification time and increased completeness of documentation of physician receipt in comparison to our prior practice and to the published literature. It has also decreased manual notification effort while providing a standardized process of reporting of critical results in compliance with The Joint Commission goals.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. Authors’ Disclosures of Potential Conflicts of Interest: No authors declared any potential conflicts of interest. Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript. Acknowledgments: We thank the managers and supervisors of the Clinical Diagnostic Laboratories and the Informatics Center for their assistance in this project.
References 1. Lundberg GD. When to panic over abnormal values. MLO Med Lab Obs 1972;4:47–54. 2. CLIA. Clinical Laboratory Improvement Amendments of 1988. Final rule, 42 CFR 405. 1992: 7001–186. 3. The Joint Commission. National Patient Safety Goals: 2010 National Patient Safety Goals. http:// www.jointcommission.org/PatientSafety/National PatientSafetyGoals (Accessed January 2010). 4. Barenfanger J, Sautter RL, Lang DL, Collins SM, Hacek DM, Peterson LR. Improving patient safety by repeating (read-back) telephone reports of critical information. Am J Clin Pathol 2004;121: 801–3. 5. Haverstick DM. Critical value called, read-back obtained. Am J Clin Pathol 2004;121:790 –1. 6. Howanitz PJ, Steindel SJ, Heard NV. Laboratory critical values policies and procedures: a College of American Pathologists Q-probes study in 623 insti-
tutions. Arch Pathol Lab Med 2002;126:663–9. 7. Valenstein PN, Wagar EA, Stankovic AK, Walsh MK, Schneider F. Notification of critical results: a College of American Pathologists Q-Probes study of 121 institutions. Arch Pathol Lab Med 2008; 132:1862–7. 8. Kuperman GJ, Teich JM, Tanasijevic MJ, Ma’Luf N, Rittenberg E, Jha A, et al. Improving response to critical laboratory results with automation: results of a randomized controlled trial. J Am Med Inform Assoc 1999;6:512–22. 9. Poon EG, Kuperman GJ, Fiskio J, Bates DW. Realtime notification of laboratory data requested by users through alphanumeric pagers. J Am Med Inform Assoc 2002;9:217–22. 10. McDowell J. Critical result reporting. AACC Clin Lab News 2007;33. 11. Kost GJ. Table of critical limits. MLO Med Lab Obs 2004;36:6 –7.
12. Dighe AS, Jones JB, Parham S, Lewandrowski KB. Survey of critical value reporting and reduction of false-positive critical value results. Arch Pathol Lab Med 2008;132:1666 –71. 13. Dighe AS, Rao A, Coakley AB, Lewandrowski KB. Analysis of laboratory critical value reporting at a large academic medical center. Am J Clin Pathol 2006;125:758 – 64. 14. Piva E, Sciacovelli L, Zaninotto M, Laposata M, Plebani M. Evaluating laboratory critical value reporting and improving communication. Am J Clin Pathol 2009;131:432– 41. 15. Hanna D, Griswold P, Leape LL, Bates DW. Communicating critical test results: safe practice recommendations. Jt Comm J Qual Patient Saf 2005; 31:68 – 80. 16. Shabot M, LoBue M, Chen J. Wireless clinical alerts for physiologic, laboratory and medication data. Proc AMIA Symp 2000:789 –93.
Clinical Chemistry 56:3 (2010) 423
