Journal of Critical Care 29 (2014) 445–449
Contents lists available at ScienceDirect
Journal of Critical Care journal homepage: www.jccjournal.org
In-hospital fellow coverage reduces communication errors in the surgical intensive care unit☆ Mallory Williams, MD, MPH a,⁎, Rodrigo F. Alban, MD b, James P. Hardy, MBBS, MD c, David A. Oxman, MD d, Edward R. Garcia, MD c, Nathanael Hevelone, MPH c, Gyorgy Frendl, MD, PhD e, f, Selwyn O. Rogers Jr., MD, MPH g a
Department of Surgery, University of Toledo Medical Center, Toledo, OH Department of Surgery, Orlando Regional Medical Center, Orlando, FL c Department of Surgery, Brigham and Women’s Hospital, Boston, MA d Department of Medicine, Thomas Jefferson University Hospital, Philadelphia, PA e Department of Anesthesia, Brigham and Women's Hospital, Boston, MA f Center for Surgery and Public Health, Brigham and Women's Hospital, Boston, MA g Department of Surgery, Temple University Hospital, Philadelphia, PA b
a r t i c l e
i n f o
a b s t r a c t
Keywords: Communication Resident Fellow Intensive Care Unit
Background: Staff coverage strategies of intensive care units (ICUs) impact clinical outcomes. High-intensity staff coverage strategies are associated with lower morbidity and mortality. Accessible clinical expertise, team work, and effective communication have all been attributed to the success of this coverage strategy. We evaluate the impact of in-hospital fellow coverage (IHFC) on improving communication of cardiorespiratory events. Methods: A prospective observational study performed in an academic tertiary care center with highintensity staff coverage. The main outcome measure was resident to fellow communication of cardiorespiratory events during IHFC vs home coverage (HC) periods. Results: Three hundred twelve cardiorespiratory events were collected in 114 surgical ICU patients in 134 study days. Complete data were available for 306 events. One hundred three communication errors occurred. IHFC was associated with significantly better communication of events compared to HC (P b .0001). Residents communicated 89% of events during IHFC vs 51% of events during HC (P b .001). Communication patterns of junior and midlevel residents were similar. Midlevel residents communicated 68% of all on-call events (87% IHFC vs 50% HC, P b .001). Junior residents communicated 66% of events (94% IHFC vs 52% HC, P b .001). Communication errors were lower in all ICUs during IHFC (P b .001). Conclusions: IHFC reduced communication errors. © 2014 Elsevier Inc. All rights reserved.
1. Introduction
role of the staff intensivist as the direct care provider in the ICU, very little is known about the role of critical care fellows in impacting overall patient care processes in the adult ICUs of academic tertiary care medical centers. Given that critical care fellows are ubiquitous in many such settings, evaluation of the impact of critical care fellows on processes and outcomes of care is important. Furthermore in an era of both accountable care and work hour restrictions, it is imperative that academic health care centers better understand which health care delivery processes produce quality outcomes and add value. The Measuring Communication in the Surgical Intensive Care Unit Study demonstrated that communication errors occurred 33% of the time between residents and fellows [5]. This prospective observational study also demonstrated that 77% of all communication errors occurred during the late shift where home coverage was more likely [5]. Bloomfield et al demonstrated that use of residents and fellows in ICU staffing at a tertiary care health center was more cost efficient than non-physician providers [6]. Furthermore, Pollack et al attributes
Intensive care unit staffing structure affects clinical outcomes [1,3]. Closed unit staffing models with an intensivist-led multidisciplinary team leads to shorter intensive care unit (ICU) lengths of hospital stay, less resource utilization, and decreased hospital and ICU mortality [1–3]. Reflecting these findings, the Society of Critical Care Medicine put forth guidelines for adult Intensive Care Units in which they recommended continuous in-hospital staffing by intensivists [4]. While most studies have focused upon the coverage model and the
☆ Supported by Brigham and Women’s Hospital Department of Surgery. ⁎ Corresponding author. Division of Trauma, Critical Care, & Acute Care Surgery, Department of Surgery, University of Toledo College of Medicine, 3000 Arlington Ave, MS 1095, Toledo, OH 43623. Tel.: + 1 419 383 6940; fax: + 1 419 383 3057. E-mail address: [email protected] (M. Williams). 0883-9441/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcrc.2014.01.004
446
M. Williams et al. / Journal of Critical Care 29 (2014) 445–449
the higher severity-adjusted mortality seen in teaching hospitals to possibly unsupervised resident patient care in the pediatric ICU [7]. When comparing similar pediatric ICUs staffed by critical care fellows and residents vs ICUs only staffed by residents, Pollack et al found that pediatric ICUs with critical care fellows had lower severity and diagnosis-adjusted mortality rates than pediatric ICUs without critical care fellows [8]. Haun speculated that the rationale for the superior outcomes of pediatric ICUs with critical care fellows was due to better resident supervision [9]. He pointed out that 6 of the 8 pediatric ICUs that had critical care fellows had 24 hour in-house fellow coverage. Also, the remaining 2 pediatric ICUs with critical care fellows had in-hospital fellow coverage 30% to 40% of nights. This in-hospital coverage created a higher level of resident supervision. The specific mechanism responsible for effectuating superior clinical outcomes has yet to be described. We hypothesize that improved resident-fellow patient care communication is responsible for these superior clinical outcomes. The impact of critical care fellows on improving patient care communication and decreasing communication errors has not been previously described. In this study we evaluate the impact of in-hospital fellow coverage (IHFC) on resident-fellow communication rates of major cardiorespiratory events. 2. Methods The institutional review board of Brigham and Women's Hospital approved this prospective observational trial. Three surgical ICUs at the Brigham and Women's Hospital were included in the study. These units included trauma/burn, general/vascular surgery, and general thoracic ICUs. The ICUs were led by boardcertified intensivists and covered by dedicated physician teams with the sole responsibility of caring for ICU patients. For the purposes of this study, an ICU study day was defined as a 24-hour period covered in an individual ICU by the resident-fellow dyad. There were 68 ICU study days in both the observational and interventional period (136 total study days). This study evaluates the prospectively collected data on communication errors from both the observational and interventional period. This study was performed using consecutive fellow coverage days of all three ICUs. Prospectively, we defined four cardiorespiratory events: hypotension, new arrhythmias, tachypnea, and desaturation. All cardiorespiratory events were required to be sustained for ≥ 2 minutes and charted on the flow sheets. Hypotension was defined as a systolic blood pressure of less than 90 mm hg or 10 mm hg below baseline. Any new arrhythmia was deemed a reportable event. Tachypnea was defined as a sustained respiratory rate greater than 35. Desaturation was defined as a pulse oximetry reading less than 90% or 10% below baseline. Events that a resident were not made aware of were not included in data collection. There were 5 critical care fellows during this study (3 from surgery and 2 from anesthesia). All of the fellows underwent a training session defining the sentinel cardiorespiratory events and short-term outcomes. Fellows participated in daily rounds and recorded the cardiorespiratory events on each patient. Events were then coded as communicated, non-communicated, or miscommunicated during both phases of the study. Non-communicated or miscommunicated events were considered communication errors. Fellow on call status during the cardiorespiratory event was recorded as in-house or home. The data collection always occurred during the morning attending rounds. All ICU nursing flow sheets were reviewed for events by the covering fellow either before or after morning attending rounds. For all non-communicated events, residents were asked if they were aware of the event. The timing of the event was a mandatory part of the question. For example, “Did you know Mrs. Jones was hypotensive at 2 am?” Any events that residents were unaware of were not included in the data collection. Judgments of the nature of the
communication (noncommunicated, miscommunicated, or communicated) were made by the fellow. Any event where the fellow was informed and later developed a significantly different understanding of the same event was considered miscommunicated. A Communication error was defined as non-communication or mis-communication of a cardiorespiratory event. At the beginning of the study, residents were given a special communication seminar with their usual orientation at the beginning of the rotation. The importance of communication as a benchmark for quality care in the ICU was formally addressed. Specific methods of communication were reviewed. Paging, cellular phone, and home phone access were prioritized. Hierarchy issues were addressed. Residents were formally invited to discuss any and all patient care issues at any time. The seminar emphasized a shift from individual, flawless performance to a relational coordination model consisting of frequent, timely, and accurate communication, mutual problem solving, shared goals, and mutual respect. Residents had 3 ways (pager, cell, or home number) to contact fellows or attending staff. 12 residents were observed during the study. There were 4 residents staffing each ICU with one resident with on-call responsibility at a time. The critical care fellows in this study take home call. However, the acuity of the surgical ICU often demands that critical care fellows perform in-hospital coverage for significant periods of time. Fellows recorded their status as in-hospital or home when they reported each event on the study data sheet. Residents in this study worked traditional shifts (24-hours). No work hours violations were reported during the study. This study was powered to detect a 20% difference in the communication rate of cardiorespiratory events between the observational and interventional arms. χ 2 Analysis was used and the expected resident to fellow communication rates were set at 90% for junior residents (PGY 1 and 2) and 80% for midlevel (PGY 3) residents. SAS software (version 8.0) was used for the statistical analysis. 3. Results There were a total of 312 cardiorespiratory events collected in 114 patients in the surgical intensive care unit over 136 study days. There were 183 (59%) events of hypotension, 47 (15%) events of new arrhythmia, 45 (14%) events of desaturation, and 37 (12%) events of tachypnea (see Fig. 1). Of the 312 cardiorespiratory events, there was fellow coverage data on 306 events. Of the 306 events there were a total of 103 communication errors collected of which complete data was available for 101. During IHFC, 89% (116/131) of all events encountered by residents were communicated (see Fig. 2). During fellow home coverage periods 50% of events were communicated (see Fig. 2). The communication patterns of junior (PGY 1 & 2) and
Events
Number of Events
Type of Communication Communicated
Communication Errors
Hypotension 183
122 (67%)
61 (33%)
Arrhythmias
47
30 (64%)
17 (36%)
Desaturations 45
30 (67%)
15 (33%)
Tachypnea
26 (70%)
11 (30%)
37
Fig. 1. Communication of cardiorespiratory events.
M. Williams et al. / Journal of Critical Care 29 (2014) 445–449
Communication of
In Hospital
Fellow
Fellow Coverage
Home Coverage
(IHFC)
(HC)
Events
Communicated
89% (116)*
50% (89)
Communication 11% (15)* Errors
50% (88)
Total Events
175
131
Fig. 2. In-hospital fellow coverage vs home coverage. ⁎IHFC vs HC groups P b .0001.
midlevel residents (PGY 3) were similar and significantly higher during IHFC (P b .001) vs HC. Midlevel residents communicated 68% of all events (87% during IHFC vs 50% during HC) (see Fig. 3). Junior residents communicated 66% of all events (94% during IHFC vs 52% during HC) (see Fig. 3). Fewer communication errors occurred in all ICUs of the study when there was in-hospital fellow coverage vs home coverage (P b .001) (see Fig. 4). Most communication errors in this study occurred in the late shift (9 pm to 7 am). During IHFC 75% of cardiorespiratory events were successfully communicated vs a 51% communication rate during HC (P b .015) (Fig. 5). While IHFC reduced communication errors of cardiovascular events short term clinical outcomes were not statistically different between IHFC and HC periods. 4. Discussion This study demonstrates that IHFC significantly improves resident to fellow communication of four key cardiorespiratory events, hypotension, new arrhythmia, desaturation, and tachypnea, meaning less communication errors occur. Residents were significantly more likely to report important changes in patient status if the fellow was in-hospital compared to being at home. Despite the fact that short term clinical outcomes were not statistically different for IHFC or HC periods, these results represent an important staffing structural observation for academic tertiary care centers. As residents work less hours it is essential to develop tools and team structures to facilitate effective patient care communication. Our study was only designed to evaluate short term improvements of cardiovascular events and therefore the impact of improved communication on long term clinical outcomes is unknown. We, however, do know that a
447
culture in the ICU (and for that matter all of graduate medical education) in which critical clinical information is interpreted by the more experienced physician is the preferential medical decisionmaking structure to effectuate optimal outcomes. Therefore improving communication of critical information from resident to fellow we believe to be a valuable quality improvement initiative. In this study, 42% (127/306) of events occurred while a critical care fellow was in hospital. The majority of these cardiorespiratory events (91%) were communicated by the resident to the fellow. However, 58% (179/306) of events occurred while fellows were at home, and only 51% were communicated. We believe these results to be reflective of a graduate medical education model that is not technologically empowered. Communication platforms such as texting, email, Face Time, Skype, and Photo Chat while preferred and mastered by many of today's trainees are actively discourage over traditional landline or cellular conversations. Newer communication methods with HIPAA compliance and feedback confirmation may improve patient care communication, resulting in more robust clinical decision making processes for critical care patients. Because of the new Accreditation Council of Graduate Medical Education resident work hour restrictions, it has become more difficult to provide 24-hour ICU in-house coverage utilizing general surgery residents. Notably, surgical interns are restricted to maximum 16-hour shifts, which relegates them to become part of a day-night float pool. Gordon et al conducted a national survey of the Association of Program Directors of Surgery [10]. This survey revealed that 37% of programs have residents other than general surgery residents providing cross-coverage and writing orders for surgical ICU patients; 24% used a night-float system in an attempt to maintain continuity of care in the surgical ICU [10]. It is difficult to argue that a night float system provides continuity of care for patients, however there are few data of surgery residents that measure ICU outcomes achieved by the system. However, a study of obstetricians, found that a night float system for attending physicians achieved a decreased use of induction of labor, a more likely use of oxytocin augmentation, and less likely to manually extract the placenta or perform episiotomy [11]. No prospective outcomes data on surgical ICU patients exists reflecting the use of these strategies. Some have suggested use of physician extenders in the surgical ICU. In the study by Gordon et al., physician extenders were used in 30% of programs to help cover the surgical ICU during the day time, and 11% used physician extenders to help cover the surgical ICU at night [10]. There are both retrospective and prospective studies in medical intensive care units that suggest physician extenders provide equal or superior care to physician based staffing models [12,13]. Outcomes data is currently lacking in surgical ICUs.
Resident
In-Hospital Fellow
Fellow
Total Communicated
Classification
Coverage
Home Coverage
Events
(IHFC)
(HC)
Junior Residents
94% (30/32) *
52% (34/65)
66% (64/97)
87% (86/99) *
50% (55/110)
68% (185/209)
89% (116/131)
51% (89/175)
67% (206/306)
(PGY 1 & 2) Midlevel Residents (PGY 3) All Residents
Fig. 3. Resident-fellow patient care communication patterns. ⁎IHFC vs HC P b .001.
448
M. Williams et al. / Journal of Critical Care 29 (2014) 445–449
ICU
In-Hospital Fellow
Fellow Home
Total
Coverage
Coverage
Communication
(IHFC)
(HC)
Trauma/Burn
96% (26/27) *
42% (26/62)
58% (52/89)
General Thoracic
86% (44/51) *
55% (27/49)
71% (71/100)
General Surgery
87% (46/53) *
56% (36/64)
70% (82/117)
Totals
89% (116/131) *
51% (89/175)
67% (205/306)
Fig. 4. Resident-fellow communication patterns by ICU. ⁎IHFC vs HC P b .001.
We also know that patient hand-offs and sign-outs provide excellent opportunities for medical errors and suboptimal continuity of patient care [14]. Data are available from medical intensive care unit studies. The Harvard Work Hours, Health and Safety Study reported a 22% higher total rate of serious errors and 5.6 times as many serious diagnostic errors for interns, while working a traditional schedule (24 hours or more) in the medical ICU [15]. Furthermore, residents working extended-duration shifts self- reported more significant medical errors, adverse events, and attention failures [16]. Residents in this study worked traditional shifts (24 hours every third to fourth night, 6 days a week). It is possible that they would be susceptible to fatigue-related decreases in performance. This may partially explain the communication failures. As stakeholders demand that we provide safer and more costeffective care, it will also be important for teaching institutions to understand the potential impact of residents and fellows in the surgical ICU. Parshuram et al surveyed Canadian adult and pediatric ICU directors to understand in-hospital, overnight physician staffing [17]. With a 98% (88/90) response rate, he found that 60% of ICUs had in-hospital physician coverage, but only 13% had in-hospital coverage by ICU attendings. The ICUs with in-hospital coverage by attendings had more ICU beds but fewer intensivists. These staff intensivists who provided in-hospital coverage had a median postgraduate experience of 3 years with 48% having less than 3 months of ICU experience. Parshuram et al. concluded that only a minority of Canadian ICUs comply with the SCCM's guidelines [17]. Angus et al also surveyed ICU directors in the United States to ascertain staffing models and Leapfrog compliance [18]. The response rate was 33.5% (393/1,173). The survey revealed that few ICUs had any in-hospital physician coverage outside of weekday daylight hours. Only 20% of ICUs had inhospital coverage during weekend days, 12% during weeknights, and 10% during weekend nights [18]. Angus' survey showed that only 4% of 255 of surveyed ICUs met the full Leapfrog standards [18]. A subsequent survey by Pronovost et al of chief medical officers and ICU directors at hospitals in 6 U.S. regions sought to understand the level of ICU physician staffing and commitment to Leapfrog standards [19]. In this survey, 61% of chief medical officers cited implementation costs
as a barrier to adopting the Leapfrog standards, and 55% of ICU directors cited controversy at the medical staff level as a barrier to implementation [19]. The aforementioned surveys detail the reality of Canada's and the United States' capacity to meet the challenge of inhospital coverage across the full spectrum of hospital ICUs. The market has attempted to adjust to a severe manpower shortage by creating telemedicine ventures designed to increase coverage of ICUs by trained intensivists. There is some evidence that telemedicine programs are safe and effective and no more costly than increased staffing protocols [20]. Breslow et al demonstrated that a supplemental remote ICU program integrated with new information technology and a delivery network improved patient outcomes [20]. Mortality, length of stay, and variable costs per case all were lower during the telemedicine ICU coverage period of the trial [20]. Residents and fellows are an essential component of the staffing structure of academic tertiary care medical centers. These trainees are clearly effectors for a range of critical care activity. Effective residentfellow coordination in the ICU is a strategy to improve processes of care and clinical outcomes. In this study, we demonstrated that the presence of in-hospital fellow coverage improves resident communication of key cardiorespiratory events. Attempts to improve overall communication in surgical ICUs on all levels may improve safety and efficiency, while decreasing costs. 5. Conclusion Fellow in-hospital call coverage reduced resident communication errors of cardiorespiratory events in the surgical ICU. Inhospital critical care fellows may be a structural buttress to improve quality and add value to critical care services in academic medical centers.
Acknowledgments We would like to thank the Mosely Professor and Chairman of Surgery, Michael J. Zinner, MD, FACS, for supporting this research. We
Type of Communication Coverage Type Communicated
Communication Errors
In-hospital Fellow Coverage
n 24
% 75% *
n 8
% 25% *
Home Coverage
71
51%
68
49%
Fig. 5. Communication patterns during the late shift (9 PM-7
AM)
for both IHFC and HC. ⁎IHFC vs HC P b .015.
M. Williams et al. / Journal of Critical Care 29 (2014) 445–449
are grateful to Claire O'Connell for providing the Database support and design for this project. We would also like to thank both the residents and the ICU nurses of the Brigham and Women's Hospital ICUs for their participation. References [1] Pronovost PJ, Angus DC, Dorman T, et al. Physician staffing patterns and clinical outcomes in critically ill patients. A systematic review. JAMA 2002;288(17): 2151–62 [PMID 12413375]. [2] Pronovost PJ, Jenckes MW, Dorman T, et al. Organizational characteristics of intensive care units related to outcomes of abdominal aortic surgery. JAMA 1999;281(14):1310–7 [PMID 10208147]. [3] Hanson III CW, Deutschman CS, Anderson II HL, et al. Effects of an organized critical care service on outcomes and resource utilization: a cohort study. Crit Care Med 1999;27:270–4. [4] Haupt MT, Bekes CE, Brilli RJ, et al. Guidelines on critical care services and personnel: Recommendations based on a system of categorization of three levels of care. Crit Care Med 2003;31(11):2677–83. [5] Williams M, Alban RF, Hardy JP, et al. Improving communication in the surgical intensive care unit. J Am Coll Surg Jan 2010;210(1):17–22. [6] Bloomfield EL, Divertie GD, Burger CD, et al. A comparison of intensive care unit physician staffing costs at the 3 Mayo clinic sites. Mayo Clin Proc 2006;81: 1457–61. [7] Pollack MM, Cuerdon TT, Patel KM, et al. Impact of quality of care factors on pediatric intensive care unit mortality. JAMA 1994;272(12):941–6. [8] Pollack MM, Patel KM, Ruttimann UE. Pediatric critical care training programs have a positive effect on pediatric intensive care mortality. Crit Care Med 1997;25 (10):1637–42.
449
[9] Haun SE. Positive impact of pediatric critical care fellows on mortality: Is it merely a function of resident supervision. Crit Care Med 1997;25(10):1622–3. [10] Gordon CR, Axelrad A, Alexander JB, et al. Care of critically ill surgical patients using the 80-hour Accreditation Council of Graduate Medical Education workweek guidelines: a survey of current strategies. Am Surg 2006;72(6):497–9. [11] Barber EL, Eisenberg DL, Grobman WA. Type of attending call schedule and changes in labor management and outcome. Obstet Gynecol Dec 2011;118(6): 1371–6. [12] Hoffman LA, Tasota FJ, Zullo TG, Scharfenberg C, Donahoe MP. Outcomes of care managed by an acute care nurse practicioner/attending physician team in a subacute medical intensive care unit. Am J Crit Care Mar 2005;14(2):121–30. [13] Gershengorn HB, Wunsch H, Wahab R, Leaf D, Brodie D, Li G, et al. Impact of nonphysician staffin on outcomes in a medical ICU. Chest 2011;139(6):1347–53. [14] Bates DW, Kaelber DC. Health information exchange and patient safety. J Biomed Inform 2007;40(6):s40–5. [15] Landrigan CP, Rothschild JM, Cronin JW, et al. Effect of reducing intern's work hours on serious medical errors in intensive care units. New Eng J Med 2004;351 (18):1838–48. [16] Barger LK, Ayas NT, Cade BT, et al. Impact of extended-duration shifts on medical errors, adverse events, and attentional failures. PLoS Med 2006;3(12):487. [17] Parshuram CS, Kirpalani H, Mehta S, et al. In-house, overnight physician staffing: a cross-sectional survey of Canadian adult and pediatric intensive care units. Crit Care Med 2006;34(6):1674–8. [18] Angus DC, Shorr AF, White A, et al. Critical care delivery in the United States: distribution of services and compliance with Leapfrog recommendations. Crit Care Med 2006;34(4):1262–3. [19] Pronovost P, Thompson DA, Holzmueller CG, et al. Impact of the Leapfrog Group’s intensive care unit physician staffing standard. J Crit Care 2007;22(2):89–96. [20] Breslow MJ, Rosenfeld BA, Doerfler M, et al. Effect of a multiple-site intensive care unit telemedicine program on clinical and economic outcomes: An alternative paradigm for intensivist staffing. Crit Care Med 2004;32:31–8.
Contents lists available at ScienceDirect
Journal of Critical Care journal homepage: www.jccjournal.org
In-hospital fellow coverage reduces communication errors in the surgical intensive care unit☆ Mallory Williams, MD, MPH a,⁎, Rodrigo F. Alban, MD b, James P. Hardy, MBBS, MD c, David A. Oxman, MD d, Edward R. Garcia, MD c, Nathanael Hevelone, MPH c, Gyorgy Frendl, MD, PhD e, f, Selwyn O. Rogers Jr., MD, MPH g a
Department of Surgery, University of Toledo Medical Center, Toledo, OH Department of Surgery, Orlando Regional Medical Center, Orlando, FL c Department of Surgery, Brigham and Women’s Hospital, Boston, MA d Department of Medicine, Thomas Jefferson University Hospital, Philadelphia, PA e Department of Anesthesia, Brigham and Women's Hospital, Boston, MA f Center for Surgery and Public Health, Brigham and Women's Hospital, Boston, MA g Department of Surgery, Temple University Hospital, Philadelphia, PA b
a r t i c l e
i n f o
a b s t r a c t
Keywords: Communication Resident Fellow Intensive Care Unit
Background: Staff coverage strategies of intensive care units (ICUs) impact clinical outcomes. High-intensity staff coverage strategies are associated with lower morbidity and mortality. Accessible clinical expertise, team work, and effective communication have all been attributed to the success of this coverage strategy. We evaluate the impact of in-hospital fellow coverage (IHFC) on improving communication of cardiorespiratory events. Methods: A prospective observational study performed in an academic tertiary care center with highintensity staff coverage. The main outcome measure was resident to fellow communication of cardiorespiratory events during IHFC vs home coverage (HC) periods. Results: Three hundred twelve cardiorespiratory events were collected in 114 surgical ICU patients in 134 study days. Complete data were available for 306 events. One hundred three communication errors occurred. IHFC was associated with significantly better communication of events compared to HC (P b .0001). Residents communicated 89% of events during IHFC vs 51% of events during HC (P b .001). Communication patterns of junior and midlevel residents were similar. Midlevel residents communicated 68% of all on-call events (87% IHFC vs 50% HC, P b .001). Junior residents communicated 66% of events (94% IHFC vs 52% HC, P b .001). Communication errors were lower in all ICUs during IHFC (P b .001). Conclusions: IHFC reduced communication errors. © 2014 Elsevier Inc. All rights reserved.
1. Introduction
role of the staff intensivist as the direct care provider in the ICU, very little is known about the role of critical care fellows in impacting overall patient care processes in the adult ICUs of academic tertiary care medical centers. Given that critical care fellows are ubiquitous in many such settings, evaluation of the impact of critical care fellows on processes and outcomes of care is important. Furthermore in an era of both accountable care and work hour restrictions, it is imperative that academic health care centers better understand which health care delivery processes produce quality outcomes and add value. The Measuring Communication in the Surgical Intensive Care Unit Study demonstrated that communication errors occurred 33% of the time between residents and fellows [5]. This prospective observational study also demonstrated that 77% of all communication errors occurred during the late shift where home coverage was more likely [5]. Bloomfield et al demonstrated that use of residents and fellows in ICU staffing at a tertiary care health center was more cost efficient than non-physician providers [6]. Furthermore, Pollack et al attributes
Intensive care unit staffing structure affects clinical outcomes [1,3]. Closed unit staffing models with an intensivist-led multidisciplinary team leads to shorter intensive care unit (ICU) lengths of hospital stay, less resource utilization, and decreased hospital and ICU mortality [1–3]. Reflecting these findings, the Society of Critical Care Medicine put forth guidelines for adult Intensive Care Units in which they recommended continuous in-hospital staffing by intensivists [4]. While most studies have focused upon the coverage model and the
☆ Supported by Brigham and Women’s Hospital Department of Surgery. ⁎ Corresponding author. Division of Trauma, Critical Care, & Acute Care Surgery, Department of Surgery, University of Toledo College of Medicine, 3000 Arlington Ave, MS 1095, Toledo, OH 43623. Tel.: + 1 419 383 6940; fax: + 1 419 383 3057. E-mail address: [email protected] (M. Williams). 0883-9441/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcrc.2014.01.004
446
M. Williams et al. / Journal of Critical Care 29 (2014) 445–449
the higher severity-adjusted mortality seen in teaching hospitals to possibly unsupervised resident patient care in the pediatric ICU [7]. When comparing similar pediatric ICUs staffed by critical care fellows and residents vs ICUs only staffed by residents, Pollack et al found that pediatric ICUs with critical care fellows had lower severity and diagnosis-adjusted mortality rates than pediatric ICUs without critical care fellows [8]. Haun speculated that the rationale for the superior outcomes of pediatric ICUs with critical care fellows was due to better resident supervision [9]. He pointed out that 6 of the 8 pediatric ICUs that had critical care fellows had 24 hour in-house fellow coverage. Also, the remaining 2 pediatric ICUs with critical care fellows had in-hospital fellow coverage 30% to 40% of nights. This in-hospital coverage created a higher level of resident supervision. The specific mechanism responsible for effectuating superior clinical outcomes has yet to be described. We hypothesize that improved resident-fellow patient care communication is responsible for these superior clinical outcomes. The impact of critical care fellows on improving patient care communication and decreasing communication errors has not been previously described. In this study we evaluate the impact of in-hospital fellow coverage (IHFC) on resident-fellow communication rates of major cardiorespiratory events. 2. Methods The institutional review board of Brigham and Women's Hospital approved this prospective observational trial. Three surgical ICUs at the Brigham and Women's Hospital were included in the study. These units included trauma/burn, general/vascular surgery, and general thoracic ICUs. The ICUs were led by boardcertified intensivists and covered by dedicated physician teams with the sole responsibility of caring for ICU patients. For the purposes of this study, an ICU study day was defined as a 24-hour period covered in an individual ICU by the resident-fellow dyad. There were 68 ICU study days in both the observational and interventional period (136 total study days). This study evaluates the prospectively collected data on communication errors from both the observational and interventional period. This study was performed using consecutive fellow coverage days of all three ICUs. Prospectively, we defined four cardiorespiratory events: hypotension, new arrhythmias, tachypnea, and desaturation. All cardiorespiratory events were required to be sustained for ≥ 2 minutes and charted on the flow sheets. Hypotension was defined as a systolic blood pressure of less than 90 mm hg or 10 mm hg below baseline. Any new arrhythmia was deemed a reportable event. Tachypnea was defined as a sustained respiratory rate greater than 35. Desaturation was defined as a pulse oximetry reading less than 90% or 10% below baseline. Events that a resident were not made aware of were not included in data collection. There were 5 critical care fellows during this study (3 from surgery and 2 from anesthesia). All of the fellows underwent a training session defining the sentinel cardiorespiratory events and short-term outcomes. Fellows participated in daily rounds and recorded the cardiorespiratory events on each patient. Events were then coded as communicated, non-communicated, or miscommunicated during both phases of the study. Non-communicated or miscommunicated events were considered communication errors. Fellow on call status during the cardiorespiratory event was recorded as in-house or home. The data collection always occurred during the morning attending rounds. All ICU nursing flow sheets were reviewed for events by the covering fellow either before or after morning attending rounds. For all non-communicated events, residents were asked if they were aware of the event. The timing of the event was a mandatory part of the question. For example, “Did you know Mrs. Jones was hypotensive at 2 am?” Any events that residents were unaware of were not included in the data collection. Judgments of the nature of the
communication (noncommunicated, miscommunicated, or communicated) were made by the fellow. Any event where the fellow was informed and later developed a significantly different understanding of the same event was considered miscommunicated. A Communication error was defined as non-communication or mis-communication of a cardiorespiratory event. At the beginning of the study, residents were given a special communication seminar with their usual orientation at the beginning of the rotation. The importance of communication as a benchmark for quality care in the ICU was formally addressed. Specific methods of communication were reviewed. Paging, cellular phone, and home phone access were prioritized. Hierarchy issues were addressed. Residents were formally invited to discuss any and all patient care issues at any time. The seminar emphasized a shift from individual, flawless performance to a relational coordination model consisting of frequent, timely, and accurate communication, mutual problem solving, shared goals, and mutual respect. Residents had 3 ways (pager, cell, or home number) to contact fellows or attending staff. 12 residents were observed during the study. There were 4 residents staffing each ICU with one resident with on-call responsibility at a time. The critical care fellows in this study take home call. However, the acuity of the surgical ICU often demands that critical care fellows perform in-hospital coverage for significant periods of time. Fellows recorded their status as in-hospital or home when they reported each event on the study data sheet. Residents in this study worked traditional shifts (24-hours). No work hours violations were reported during the study. This study was powered to detect a 20% difference in the communication rate of cardiorespiratory events between the observational and interventional arms. χ 2 Analysis was used and the expected resident to fellow communication rates were set at 90% for junior residents (PGY 1 and 2) and 80% for midlevel (PGY 3) residents. SAS software (version 8.0) was used for the statistical analysis. 3. Results There were a total of 312 cardiorespiratory events collected in 114 patients in the surgical intensive care unit over 136 study days. There were 183 (59%) events of hypotension, 47 (15%) events of new arrhythmia, 45 (14%) events of desaturation, and 37 (12%) events of tachypnea (see Fig. 1). Of the 312 cardiorespiratory events, there was fellow coverage data on 306 events. Of the 306 events there were a total of 103 communication errors collected of which complete data was available for 101. During IHFC, 89% (116/131) of all events encountered by residents were communicated (see Fig. 2). During fellow home coverage periods 50% of events were communicated (see Fig. 2). The communication patterns of junior (PGY 1 & 2) and
Events
Number of Events
Type of Communication Communicated
Communication Errors
Hypotension 183
122 (67%)
61 (33%)
Arrhythmias
47
30 (64%)
17 (36%)
Desaturations 45
30 (67%)
15 (33%)
Tachypnea
26 (70%)
11 (30%)
37
Fig. 1. Communication of cardiorespiratory events.
M. Williams et al. / Journal of Critical Care 29 (2014) 445–449
Communication of
In Hospital
Fellow
Fellow Coverage
Home Coverage
(IHFC)
(HC)
Events
Communicated
89% (116)*
50% (89)
Communication 11% (15)* Errors
50% (88)
Total Events
175
131
Fig. 2. In-hospital fellow coverage vs home coverage. ⁎IHFC vs HC groups P b .0001.
midlevel residents (PGY 3) were similar and significantly higher during IHFC (P b .001) vs HC. Midlevel residents communicated 68% of all events (87% during IHFC vs 50% during HC) (see Fig. 3). Junior residents communicated 66% of all events (94% during IHFC vs 52% during HC) (see Fig. 3). Fewer communication errors occurred in all ICUs of the study when there was in-hospital fellow coverage vs home coverage (P b .001) (see Fig. 4). Most communication errors in this study occurred in the late shift (9 pm to 7 am). During IHFC 75% of cardiorespiratory events were successfully communicated vs a 51% communication rate during HC (P b .015) (Fig. 5). While IHFC reduced communication errors of cardiovascular events short term clinical outcomes were not statistically different between IHFC and HC periods. 4. Discussion This study demonstrates that IHFC significantly improves resident to fellow communication of four key cardiorespiratory events, hypotension, new arrhythmia, desaturation, and tachypnea, meaning less communication errors occur. Residents were significantly more likely to report important changes in patient status if the fellow was in-hospital compared to being at home. Despite the fact that short term clinical outcomes were not statistically different for IHFC or HC periods, these results represent an important staffing structural observation for academic tertiary care centers. As residents work less hours it is essential to develop tools and team structures to facilitate effective patient care communication. Our study was only designed to evaluate short term improvements of cardiovascular events and therefore the impact of improved communication on long term clinical outcomes is unknown. We, however, do know that a
447
culture in the ICU (and for that matter all of graduate medical education) in which critical clinical information is interpreted by the more experienced physician is the preferential medical decisionmaking structure to effectuate optimal outcomes. Therefore improving communication of critical information from resident to fellow we believe to be a valuable quality improvement initiative. In this study, 42% (127/306) of events occurred while a critical care fellow was in hospital. The majority of these cardiorespiratory events (91%) were communicated by the resident to the fellow. However, 58% (179/306) of events occurred while fellows were at home, and only 51% were communicated. We believe these results to be reflective of a graduate medical education model that is not technologically empowered. Communication platforms such as texting, email, Face Time, Skype, and Photo Chat while preferred and mastered by many of today's trainees are actively discourage over traditional landline or cellular conversations. Newer communication methods with HIPAA compliance and feedback confirmation may improve patient care communication, resulting in more robust clinical decision making processes for critical care patients. Because of the new Accreditation Council of Graduate Medical Education resident work hour restrictions, it has become more difficult to provide 24-hour ICU in-house coverage utilizing general surgery residents. Notably, surgical interns are restricted to maximum 16-hour shifts, which relegates them to become part of a day-night float pool. Gordon et al conducted a national survey of the Association of Program Directors of Surgery [10]. This survey revealed that 37% of programs have residents other than general surgery residents providing cross-coverage and writing orders for surgical ICU patients; 24% used a night-float system in an attempt to maintain continuity of care in the surgical ICU [10]. It is difficult to argue that a night float system provides continuity of care for patients, however there are few data of surgery residents that measure ICU outcomes achieved by the system. However, a study of obstetricians, found that a night float system for attending physicians achieved a decreased use of induction of labor, a more likely use of oxytocin augmentation, and less likely to manually extract the placenta or perform episiotomy [11]. No prospective outcomes data on surgical ICU patients exists reflecting the use of these strategies. Some have suggested use of physician extenders in the surgical ICU. In the study by Gordon et al., physician extenders were used in 30% of programs to help cover the surgical ICU during the day time, and 11% used physician extenders to help cover the surgical ICU at night [10]. There are both retrospective and prospective studies in medical intensive care units that suggest physician extenders provide equal or superior care to physician based staffing models [12,13]. Outcomes data is currently lacking in surgical ICUs.
Resident
In-Hospital Fellow
Fellow
Total Communicated
Classification
Coverage
Home Coverage
Events
(IHFC)
(HC)
Junior Residents
94% (30/32) *
52% (34/65)
66% (64/97)
87% (86/99) *
50% (55/110)
68% (185/209)
89% (116/131)
51% (89/175)
67% (206/306)
(PGY 1 & 2) Midlevel Residents (PGY 3) All Residents
Fig. 3. Resident-fellow patient care communication patterns. ⁎IHFC vs HC P b .001.
448
M. Williams et al. / Journal of Critical Care 29 (2014) 445–449
ICU
In-Hospital Fellow
Fellow Home
Total
Coverage
Coverage
Communication
(IHFC)
(HC)
Trauma/Burn
96% (26/27) *
42% (26/62)
58% (52/89)
General Thoracic
86% (44/51) *
55% (27/49)
71% (71/100)
General Surgery
87% (46/53) *
56% (36/64)
70% (82/117)
Totals
89% (116/131) *
51% (89/175)
67% (205/306)
Fig. 4. Resident-fellow communication patterns by ICU. ⁎IHFC vs HC P b .001.
We also know that patient hand-offs and sign-outs provide excellent opportunities for medical errors and suboptimal continuity of patient care [14]. Data are available from medical intensive care unit studies. The Harvard Work Hours, Health and Safety Study reported a 22% higher total rate of serious errors and 5.6 times as many serious diagnostic errors for interns, while working a traditional schedule (24 hours or more) in the medical ICU [15]. Furthermore, residents working extended-duration shifts self- reported more significant medical errors, adverse events, and attention failures [16]. Residents in this study worked traditional shifts (24 hours every third to fourth night, 6 days a week). It is possible that they would be susceptible to fatigue-related decreases in performance. This may partially explain the communication failures. As stakeholders demand that we provide safer and more costeffective care, it will also be important for teaching institutions to understand the potential impact of residents and fellows in the surgical ICU. Parshuram et al surveyed Canadian adult and pediatric ICU directors to understand in-hospital, overnight physician staffing [17]. With a 98% (88/90) response rate, he found that 60% of ICUs had in-hospital physician coverage, but only 13% had in-hospital coverage by ICU attendings. The ICUs with in-hospital coverage by attendings had more ICU beds but fewer intensivists. These staff intensivists who provided in-hospital coverage had a median postgraduate experience of 3 years with 48% having less than 3 months of ICU experience. Parshuram et al. concluded that only a minority of Canadian ICUs comply with the SCCM's guidelines [17]. Angus et al also surveyed ICU directors in the United States to ascertain staffing models and Leapfrog compliance [18]. The response rate was 33.5% (393/1,173). The survey revealed that few ICUs had any in-hospital physician coverage outside of weekday daylight hours. Only 20% of ICUs had inhospital coverage during weekend days, 12% during weeknights, and 10% during weekend nights [18]. Angus' survey showed that only 4% of 255 of surveyed ICUs met the full Leapfrog standards [18]. A subsequent survey by Pronovost et al of chief medical officers and ICU directors at hospitals in 6 U.S. regions sought to understand the level of ICU physician staffing and commitment to Leapfrog standards [19]. In this survey, 61% of chief medical officers cited implementation costs
as a barrier to adopting the Leapfrog standards, and 55% of ICU directors cited controversy at the medical staff level as a barrier to implementation [19]. The aforementioned surveys detail the reality of Canada's and the United States' capacity to meet the challenge of inhospital coverage across the full spectrum of hospital ICUs. The market has attempted to adjust to a severe manpower shortage by creating telemedicine ventures designed to increase coverage of ICUs by trained intensivists. There is some evidence that telemedicine programs are safe and effective and no more costly than increased staffing protocols [20]. Breslow et al demonstrated that a supplemental remote ICU program integrated with new information technology and a delivery network improved patient outcomes [20]. Mortality, length of stay, and variable costs per case all were lower during the telemedicine ICU coverage period of the trial [20]. Residents and fellows are an essential component of the staffing structure of academic tertiary care medical centers. These trainees are clearly effectors for a range of critical care activity. Effective residentfellow coordination in the ICU is a strategy to improve processes of care and clinical outcomes. In this study, we demonstrated that the presence of in-hospital fellow coverage improves resident communication of key cardiorespiratory events. Attempts to improve overall communication in surgical ICUs on all levels may improve safety and efficiency, while decreasing costs. 5. Conclusion Fellow in-hospital call coverage reduced resident communication errors of cardiorespiratory events in the surgical ICU. Inhospital critical care fellows may be a structural buttress to improve quality and add value to critical care services in academic medical centers.
Acknowledgments We would like to thank the Mosely Professor and Chairman of Surgery, Michael J. Zinner, MD, FACS, for supporting this research. We
Type of Communication Coverage Type Communicated
Communication Errors
In-hospital Fellow Coverage
n 24
% 75% *
n 8
% 25% *
Home Coverage
71
51%
68
49%
Fig. 5. Communication patterns during the late shift (9 PM-7
AM)
for both IHFC and HC. ⁎IHFC vs HC P b .015.
M. Williams et al. / Journal of Critical Care 29 (2014) 445–449
are grateful to Claire O'Connell for providing the Database support and design for this project. We would also like to thank both the residents and the ICU nurses of the Brigham and Women's Hospital ICUs for their participation. References [1] Pronovost PJ, Angus DC, Dorman T, et al. Physician staffing patterns and clinical outcomes in critically ill patients. A systematic review. JAMA 2002;288(17): 2151–62 [PMID 12413375]. [2] Pronovost PJ, Jenckes MW, Dorman T, et al. Organizational characteristics of intensive care units related to outcomes of abdominal aortic surgery. JAMA 1999;281(14):1310–7 [PMID 10208147]. [3] Hanson III CW, Deutschman CS, Anderson II HL, et al. Effects of an organized critical care service on outcomes and resource utilization: a cohort study. Crit Care Med 1999;27:270–4. [4] Haupt MT, Bekes CE, Brilli RJ, et al. Guidelines on critical care services and personnel: Recommendations based on a system of categorization of three levels of care. Crit Care Med 2003;31(11):2677–83. [5] Williams M, Alban RF, Hardy JP, et al. Improving communication in the surgical intensive care unit. J Am Coll Surg Jan 2010;210(1):17–22. [6] Bloomfield EL, Divertie GD, Burger CD, et al. A comparison of intensive care unit physician staffing costs at the 3 Mayo clinic sites. Mayo Clin Proc 2006;81: 1457–61. [7] Pollack MM, Cuerdon TT, Patel KM, et al. Impact of quality of care factors on pediatric intensive care unit mortality. JAMA 1994;272(12):941–6. [8] Pollack MM, Patel KM, Ruttimann UE. Pediatric critical care training programs have a positive effect on pediatric intensive care mortality. Crit Care Med 1997;25 (10):1637–42.
449
[9] Haun SE. Positive impact of pediatric critical care fellows on mortality: Is it merely a function of resident supervision. Crit Care Med 1997;25(10):1622–3. [10] Gordon CR, Axelrad A, Alexander JB, et al. Care of critically ill surgical patients using the 80-hour Accreditation Council of Graduate Medical Education workweek guidelines: a survey of current strategies. Am Surg 2006;72(6):497–9. [11] Barber EL, Eisenberg DL, Grobman WA. Type of attending call schedule and changes in labor management and outcome. Obstet Gynecol Dec 2011;118(6): 1371–6. [12] Hoffman LA, Tasota FJ, Zullo TG, Scharfenberg C, Donahoe MP. Outcomes of care managed by an acute care nurse practicioner/attending physician team in a subacute medical intensive care unit. Am J Crit Care Mar 2005;14(2):121–30. [13] Gershengorn HB, Wunsch H, Wahab R, Leaf D, Brodie D, Li G, et al. Impact of nonphysician staffin on outcomes in a medical ICU. Chest 2011;139(6):1347–53. [14] Bates DW, Kaelber DC. Health information exchange and patient safety. J Biomed Inform 2007;40(6):s40–5. [15] Landrigan CP, Rothschild JM, Cronin JW, et al. Effect of reducing intern's work hours on serious medical errors in intensive care units. New Eng J Med 2004;351 (18):1838–48. [16] Barger LK, Ayas NT, Cade BT, et al. Impact of extended-duration shifts on medical errors, adverse events, and attentional failures. PLoS Med 2006;3(12):487. [17] Parshuram CS, Kirpalani H, Mehta S, et al. In-house, overnight physician staffing: a cross-sectional survey of Canadian adult and pediatric intensive care units. Crit Care Med 2006;34(6):1674–8. [18] Angus DC, Shorr AF, White A, et al. Critical care delivery in the United States: distribution of services and compliance with Leapfrog recommendations. Crit Care Med 2006;34(4):1262–3. [19] Pronovost P, Thompson DA, Holzmueller CG, et al. Impact of the Leapfrog Group’s intensive care unit physician staffing standard. J Crit Care 2007;22(2):89–96. [20] Breslow MJ, Rosenfeld BA, Doerfler M, et al. Effect of a multiple-site intensive care unit telemedicine program on clinical and economic outcomes: An alternative paradigm for intensivist staffing. Crit Care Med 2004;32:31–8.
