Share this article on:

Introduction of a Rapid Response System at a United States Veterans Affairs Hospital Reduced Cardiac Arrests

Lighthall, Geoffrey K. MD, PhD*,†; Parast, Layla M. MS; Rapoport, Lisa MD, MS§; Wagner, Todd H. PhD∥,¶

doi: 10.1213/ANE.0b013e3181e9c3f3
Critical Care, Trauma, and Resuscitation: Research Reports
Chinese Language Editions

BACKGROUND: We sought to determine the impact of a rapid response system on cardiac arrest rates and mortality in a United States veteran population.

METHODS: We describe a prospective analysis of cardiac arrests in 9 months before and 27 months after institution of a rapid response system, and retrospective analysis of mortality 3.5 years before the intervention and 27 months after the intervention. The study included all inpatients from a university-affiliated United States Veterans Affairs Medical Center, before and after implementation of a rapid response system, including an educational program, patient calling criteria, and a physician-led medical emergency team. Primary end points were hospital-wide cardiac arrests and mortality rates normalized to hospital discharges. Comparisons of event rates between various time points during the implementation process were made by analysis of variance.

RESULTS: Three hundred seventy-eight calls were made to the medical emergency team in the time period studied. Compared with preintervention time points, cardiac arrests were reduced by 57%, amounting to a reduction of 5.6 cardiac arrests per 1000 hospital discharges (P < 0.01). Mortality was reduced during the intervention, but this was attributable to a natural decrease occurring over all phases of the study.

CONCLUSIONS: A significant reduction in the rate of cardiac arrests was realized with this intervention, as well as a trend toward lower mortality. We estimate that 51 arrests were prevented in the timeframe studied. Our results suggest that further reductions in morbidity can be realized by expansion of rapid response systems throughout the Veterans Affairs network.

Published ahead of print July 12, 2010 Supplemental Digital Content is available in the text.

From the *Department of Anesthesia, §Critical Care Medicine, and VA Health Economics Resource Center, Stanford University School of Medicine, Stanford, California; Department of Anesthesia, and VA Health Economics Resource Center, Veterans Affairs Medical Center, Palo Alto, California; and Department of Biostatistics, Harvard University, Boston, Massachusetts.

Disclosure: The authors report no conflicts of interest.

Reprints will not be available from the author.

Address correspondence to Geoffrey K. Lighthall, MD, PhD, Department of Anesthesiology, MC 112A, Veterans Affairs Medical Center, Palo Alto, CA 94304. Address e-mail to geoffL@stanford.edu.

Accepted May 9, 2010

Published ahead of print July 12, 2010

Rapid response systems (RRSs) enhance traditional medical care by providing additional help in evaluating and treating potentially unstable patients. The benefits derived from RRSs include reduced rates of cardiac arrest, decreased mortality, prevention of organ dysfunction, and earlier institution of important therapies.114 Some have questioned the generalizability of results from single-center studies and have cautioned against the widespread implementation of rapid response teams on that basis.1517 Given a rather uniform composition among hospitalized patients served by the United States (US) Veterans Affairs (VA) Network, interventions that produce a benefit one hospital have the potential of benefiting the wider population of patients within the system. We were therefore interested to know whether an RRS could be implemented at a VA hospital, and whether it could improve important patient outcomes.

The US Department of VA Health Care System is the nation's largest integrated health care provider with a nationwide network of more than 100 hospitals that provide acute medical and surgical care, as well as longer-term nursing and hospice care. The VA is an integral part of the nation's safety net.18 Approximately two-thirds of the VA users receive free services because of a service-connected disability or their income status, and many veterans present with advanced illness and comorbid chronic conditions. Implementation of the medical emergency team was based on the hypothesis that the delivery of extra personnel and resources based on the dissemination of written criteria would have a beneficial impact on cardiac arrest rates and other forms of morbidity in non–intensive care unit (ICU) patients. An examination of cardiac arrest and mortality rates before and after this intervention is presented here. This is the first study we are aware of to describe the implementation and results of an RRS in a US VA Hospital.

Back to Top | Article Outline

METHODS

The study was conducted under the approval of the Human Subjects Board at Stanford University and the Research and Development Committee at the VA Palo Alto.

Back to Top | Article Outline

Setting

The VA Palo Alto is a tertiary medical center that serves a geographical area of 13,500 square miles. There are 150 acute care beds, including 15 critical care and 15 intermediate care beds, and 60 nursing rehabilitation beds. There are approximately 13,000 annual admissions to the emergency department, with approximately 4500 acute medical/surgical admissions to the hospital each year, and more than 800 admissions to the ICU. The Palo Alto VA is affiliated with Stanford University, and residents rotate through the Palo Alto VA. The educational mission includes training of house staff from accredited residency and fellowship programs in anesthesiology, internal medicine, psychiatry, neurology, rehabilitation medicine, surgery, and a number of subspecialties. The ICU is staffed 24 hours a day by physicians who are board certified in intensive care.

Back to Top | Article Outline

RRS Implementation

A high capability medical emergency team, which was called the “eTeam” in the hospital, began operation in late June 2005; the composition of the team is described in Table 1. In mid-2004, a steering committee consisting of major academic and clinical service departments and patient safety experts was assembled to establish variables of operation, calling criteria, team composition, and barriers to implementation and success. An educational/in-service program directed at ward nurses and medical and surgical departments was launched 4 months before full operation of the system. Posters containing calling criteria (Table 2) were placed throughout the hospital, and identification badge holder inserts containing the criteria were given to all nursing, medical, and ancillary personnel. Also in 2004, the cardiopulmonary resuscitation committee began investigating antecedents to cardiac arrest calls; this pre-event and event analysis effort was later expanded to analyze all RRSs and arrest calls. Meetings of both the eTeam steering committee and cardiopulmonary resuscitation committee include physician and nursing leaders corresponding to the wards and areas of the intervention. Meetings centered on discussion of event rates and team and provider conduct, efficacy of intervention, and patient outcomes. Sharing topics discussed with all applicable trainees and staff was encouraged.

Table 1

Table 1

Table 2

Table 2

Specific materials brought by the eTeam to a patient's bedside are also noted in Table 1. The eTeam is paged via a beeper system to reduce the crowds often attracted by overhead announcements. Hospital personnel are directed to summon the primary caregiving team at the same time as paging the eTeam.

Back to Top | Article Outline

Data Sources

Cardiac arrest calls were recorded by the hospital operators and on paper event records kept by the arrest team pharmacist. For purposes of this analysis, cardiac arrests were defined as episodes of pulseless ventricular tachycardia, ventricular fibrillation, symptomatic bradycardia or supraventricular tachycardia, pulseless electrical activity, asystole, and respiratory failure requiring immediate tracheal intubation. Cardiac arrests in areas not served by the eTeam (areas out of the main hospital such as the psychiatry hospital, outpatient clinics, and cafeteria) were not included in the analysis on eTeam effects on arrests. Likewise, cardiac arrest events in the emergency department were not included in the analysis of events, because arrests at this location would not likely be affected by the presence of a medical emergency team.

Mortality data were obtained from the decedent affairs office and organized according to the patients' location at the time of death as well as resuscitation (do not resuscitate, DNR) status. Records were separated further according to hospital locations. If records indicated that a patient received emergency cardiovascular care despite a DNR order, and there was no evidence that resuscitation was halted because of DNR status, the patient was classified as a “full-code” patient.

To risk-adjust the data, we linked patient information to the VA's medical record. We extracted international statistical classification of diseases (ICD)-9 diagnostic information and used this information to compute the Deyo-modified Charlson Comorbidity Index (CCI).19 The CCI score, which is strongly associated with mortality, is based on the ICD-9 diagnoses, and the score ranges from healthiest (1) to most severe (3).20

Back to Top | Article Outline

Data Analysis

Cardiac arrest rates were organized according to ward and month and were then pooled and normalized to the total number of medical and surgical admissions for each month (events per 1000 discharges); this discharge total includes all acute admissions for medical, surgical, and neurologic conditions. Arrest rates from individual wards were normalized to monthly occupancy data to make interward comparisons, because discharge data for individual wards were not available.

Before implementation of the emergency team, the cardiac arrest team, if summoned, attended to cardiac arrests, near arrests, and respiratory insufficiency and other emergencies. An “airway pager” carried by the anesthesiologist on call could be activated by physicians and ward staff for airway and respiratory emergencies. There were no guidelines or criteria for summoning the arrest team or anesthesiologist before implementation of the eTeam. A comparison of cardiac arrest rates after implementation of the RRS was compared with a control period of the 9 months before implementation, and analyzed in 9-month blocks after the start of operation. An evaluation of cardiac arrests was also performed for a chronic nursing/rehabilitation ward within the hospital that was served by the RRS; occupancy and discharge data specific to this unit were used to analyze the RRS impact on this ward. Arrest rates for time periods before and after implementation of the RRS were compared by Student t test, and by analysis of variance for analysis of multiple time periods.

Mortality rates were calculated as the number of deaths per 100 discharges from the acute care hospital. This denominator includes deaths in all acute care settings including the emergency department, ICU, and regular wards, but does not include mortality from hospice and nursing units housed within the building. Further categorization was made according to the resuscitation status of the patient, and is described in the Results section as either “all mortality” or “full-code mortality.”

Mortality data were initially analyzed using a t test. To account for possible trends in time and seasonality, the data were also analyzed as time series data. A seasonal dummy model was fit and an F test was performed to determine the need to retain seasonal effects in the model. The intervention was coded as a step intervention, i.e., with value 0 before July 2005 and value 1 July 2005 and after. The possibilities of either an immediate change or a gradual change were both evaluated in the model. The Durbin-Watson statistic was used to test for autocorrelation. Results were adjusted for possible changes in CCI score over time. This analysis was performed using R software, version 2.9.0.

Back to Top | Article Outline

RESULTS

In July 2005, an RRS featuring a physician-led medical eTeam was implemented at a hospital within the Palo Alto VA Health Care System. Through September 2007, the eTeam received 378 calls; call volume ranged between 20 and 60 calls per 1000 discharges. Respiratory insufficiency accounted for one-third of the calls, with hypotension (22%) and deterioration in mental status (20%) completing the top 3 reasons (Table 3). Notably, of the cases in which a change in mental status was stated as the primary reason to call, other objective calling criteria such as hypotension and tachycardia were present in >90% of such instances. Likewise, one of the objective criteria was present in nearly all cases in which the eTeam was activated over “concern about patient.” The “concern” category is not listed in Table 3; rather, the primary objective criteria present at the time are listed. Patients were transferred to higher levels of care in 58% of instances, and the average time per call was approximately 30 minutes. Activity of the eTeam varied according to time of day, with a daytime call volume (8 AM to 6 PM) approximately twice that of the night (Fig. 1).

Table 3

Table 3

Figure 1

Figure 1

Demographics characteristics of the sample are presented in Table 4 . Compared to after the implementation of the eTeam, the inpatient population before the implementation of the eTeam had a lower comorbidity burden (CCI preintervention 1.75 ± 0.095 [SEM], vs 1.96 ± 0.178; P < 0.001), and a slightly lower mean age (65.26 ± 0.08 years vs 65.56 ± 0.08 years; P = 0.016). Hospice mortality was also lower in the study period compared with before the study period (0.25 ± 0.03 deaths per 100 discharges vs 0.16 ± 0.02; P < 0.01). The patients were 96% male at all time periods analyzed, which is consistent with the overall population that uses the VA.

Table 4

Table 4

Interventions performed by the team are also listed in Table 3. A majority of patients received an oxygen mask (68%), arterial blood gas (67%), fluids (60%), and an electrocardiogram (56%). Tracheal intubation (13%), noninvasive ventilation (19%), and insertion of an IV line (21%) were less often performed. Very few calls (2.6%) resulted in a change of care status to comfort only or do not escalate or DNR.

After implementation of the RRS, hospital-wide mortality was 17.3% lower, with 2.24 ± 0.87 (mean ± SD) deaths per 100 discharges versus 2.71 ± 0.98 in the preintervention period (P = 0.0431) (Fig. 2). Full-code mortality was reduced 43% with a preintervention mortality of 0.68 ± 0.478 deaths per 100 discharges, versus 0.39 ± 0.292 after the RRS (P = 0.0031). Full-code and DNR mortality were divided further into ICU and ward categories. The ward data showed a sharp nonsignificant decrease in full-code mortality, whereas there was no clear trend in ward DNR mortality, or in either DNR or full-code mortality in the ICU (data not shown).

Figure 2

Figure 2

Further analyses were conducted to evaluate differences attributable to the RRS versus periodic or secular variations. Time series analysis indicated a slight decreasing trend in (all death) mortality both pre- and postintervention; this was estimated to be a decrease of 0.02244 deaths per 100 discharges per month after adjusting for seasonality (P = 0.0351). There was no evidence of an immediate or gradual change in mortality attributable to the intervention (P = 0.3981). Time series analysis also indicated a decreasing trend in full-code mortality; this was estimated to be a decrease of 0.013 deaths per 100 discharges per month after adjusting for seasonality (P = 0.0041). There was no evidence of an immediate or gradual change in full-code mortality attributable to the intervention (P = 0.331). After adjusting for CCI score, there was no longer a significant decreasing trend over time in either mortality or full-code mortality.

Monthly rates of eTeam calls and cardiac arrests are presented in Figure 3. Comparing the incidence of cardiac arrest in the 9 months before team operation to the period after, there was a significant (57%) reduction in arrests across the hospital (mean of 10.1 ± 2.0 [SEM] before, vs 4.36 ± 0.9 arrests per 1000 discharges after the RRS; P < 0.01). The number of respiratory arrests was strikingly lower after the eTeam (17 in the previous year, compared with 1, 2, and 2 arrests per year in the years after the eTeam). Most cardiac arrests occurred on the acute medical/surgical wards, consisting of medical, neurologic, and surgical patients, which accounted for approximately 55% of the patients studied. When these wards are considered as a separate group, a similar 48% reduction in arrest rate is apparent (7.78 ± 1.7 to 3.80 ± 0.8 events per 1000 discharges after the RRS; P = 0.03; Fig. 3).

Figure 3

Figure 3

Further analysis according to medical and surgical patient subpopulations indicated larger arrest reductions in the latter (43% vs 64%, respectively; Table 5). Ward-level comparisons of arrests normalized to bed occupancy also revealed some heterogeneity in results, with statistically significant reductions present only in the intermediate ICU (IICU) (Table 5). Comparing the pre-RRS period with the three 9-month blocks after the intervention indicates that on a time-interval basis, a reduction in arrests was not statistically significant until months 10 to 18 after the RRS for the whole hospital, and not until months 19 to 27 for the acute medical/surgical wards (Fig. 4).

Table 5

Table 5

Figure 4

Figure 4

Back to Top | Article Outline

DISCUSSION

The adoption of an RRS was associated with a reduction of cardiac arrests and a trend toward reduced mortality. The data presented here add to the collection of studies demonstrating decreased cardiac arrest rates after the adoption of an RRS. The reduction in arrests was not observed immediately; rather, the findings were most prominent in the second year of the eTeam. The lack of clear outcome benefit in the first several months after the start of the RRS operation is consistent with the findings of others,5,6 and probably reflects the normal course of this intervention. The need for a longer run-in period may also help explain why a large Australian/New Zealand collaborative study was unable to demonstrate a reduction in cardiac arrests at an end point of 6 months.21

Despite a decrease in cardiac arrests after the eTeam was initiated, the fraction of arrest victims exhibiting one of the calling criteria before arrest remained essentially unchanged. The latter, along with the fact that approximately two-thirds of calls required transfer to either the ICU or IICU, as opposed to the 20%–45% range documented in most other studies,2226 suggests that the system was not used to its fullest potential, and that additional patient benefit could be realized with additional use.27 The reasons for what we perceive as underutilization are unclear but, from periodic review of cases, the reasons seem to include physician reluctance to accept help, refresher training of nursing and ancillary staff is too infrequent, and poor prior experiences with team members.

The diurnal variation in call volume has been observed in other studies of RRSs and has been interpreted as a sign of underutilization or poorer oversight of patients during the night.8,9 Staff behavior such as entry to rooms for medicine and meal distribution, physical therapy, patient education, and transfer of other patients probably represents a subtle form of monitoring that is not present at night, and that may contribute to greater daytime detection of events. Despite an increased presence of physicians and other staff during the day, a higher daytime call volume may also result from periods of resident unavailability or because response times are slow because of rounds, surgery, educational activities, or family meetings. Nights have also been associated with delays in defibrillation, and a lower overall survival from in-hospital cardiac arrest in an analysis of a large voluntary database.2830 The frequency of acute cardiac neurologic and respiratory disease is subject to diurnal variation, with a greater severity recorded in late night to early morning.31 Together, these findings indicate a role for improved vigilance, monitoring, and perhaps change of practice during night hours. The data also suggest the possibility that the benefit of an RRS is based more on monitoring and event detection than the composition of the response team.

The points raised above, that staff behavior and training as well as level of monitoring may influence call frequency, is supported by the finding of a higher call volume in our IICU (Table 3). All patients in this unit have continuous electrocardiogram and oximetry; the unit has an open floor plan with a greater physician and nurse presence. Additionally, the IICU nurses respond to eTeam calls and are therefore likely to have the greatest understanding of the rationale, function, and benefits of the RRS.

Even with what we interpret to be incomplete use of the RRS, benefits were realized. Regression and time series analyses were used to assess the contribution of seasonal effects and other trends over the time studied; these analyses did reveal a natural decrease in mortality of both full-code and all-type mortality that questions the overall impact of the RRS on mortality. We also considered whether a change in patient characteristics could affect our results. Computation of monthly pooled comorbidity indices before and throughout the study period showed that our results were obtained at a time when the population of inpatients was actually becoming older, harboring a higher burden of chronic disease, and therefore likely to have a higher mortality. Diversion of patients toward palliative care was unlikely to have had a positive impact on hospital mortality, because the hospice unit mortality rate actually decreased during the study period. Likewise, it is reasonable to consider whether the movement of unstable patients from the ward to the ICU shifted the arrest and mortality burden to the latter location.28 We could not identify any trend in the data to support the latter possibility, although future consideration of this question with better tracking of ICU codes is certainly warranted. Other quality-improvement measures such as the introduction of a hospitalist service, enhanced infection control practices, a ventilator-associated pneumonia prevention program, and improvements in glycemic control may have influenced the results obtained; however, all such interventions were in place for at least 10 months before the start of the eTeam.

The eTeam is a largely social intervention with multiple components whose individual contributions to patient safety and survival are not known.32 It has been encouraged to think of medical emergency teams such as ours as being part of a larger RRS that includes 4 components: (1) an afferent arm, consisting of establishment of calling criteria, education, and support systems that lead to identification of patients in need; (2) an efferent arm consisting of the emergency team and its activities; (3) administrative and governance functions; and (4) data collection and quality-improvement activities.33 Although the most visible component of the eTeam and other RRSs may be the team and its activities, current data have failed to link either success or failure to team design per se. The rate-limiting step in team involvement is its activation by the ward staff (the afferent arm), where timeliness of activation has been associated with a lower mortality for some conditions.34 Establishing the afferent arm as part of a larger system involves the deliberate education of clinicians on the nature of critical illness, the need for early intervention, and also the establishment of criteria for summoning help. The presence of the latter has, in our experience, “given permission” to ward staff to call for help with patients for whom there has been suspicion of clinical decline, but there is some ambiguity as to the true problem. Likewise, knowing that an ICU outreach team will be called or that event rates are being scrutinized, may have provided either subconscious (Hawthorne effect) or conscious motivation for ward physicians to respond in a timelier manner to calls for evaluation or assistance. One can thus envision that an increase in ward nurse and staff awareness of clinical deterioration and earlier calling of both eTeam and non-eTeam physicians may be a strong contributor to the benefits of the RRS seen here and elsewhere.

If the overall benefit of the eTeam is the system per se, rather than the team, one may question the more costly diagnostic resources used by our team, and the more advanced level of training embodied by our team and others.6,35 An understanding that intensive care began too late in many cases, combined with sporadic lack of ICU bed availability, drove us to adopt a higher level of capability that allows many aspects of intensive care, such as point-of-care blood analysis, invasive line placement, continuous display of invasive and noninvasive waveforms, medication administration, and respiratory support, to begin before arrival in the ICU.

Because some departments expressed a fear that the eTeam would deprive their trainees of valuable educational experiences, our team design embraced an educational mission, a mission based on having the most experienced personnel available present during patient care. The presence of senior ICU personnel during the day has allowed some instruction to accompany bedside care during team involvement. The presence of point-of-care blood analysis and continuous monitoring enhances the educational and safety functions of the team by providing mentorship in the use and interpretation of diagnostic studies in an acute setting.

Criticism of prior work on RRS outcomes includes a reliance on subjective outcomes, failure to report cointerventions and staffing models, poorly defined criteria for cardiac arrest calls, and underreporting of mortality postintervention by excluding patients who were made DRN/no further resuscitation by the team.15 We attempted to present all variables that may have affected the results. Cardiac arrest events were classified as such in conformance with accepted standards, and consistent criteria were used to define these events through all periods studied.36 Whole-hospital mortality included all patients in a palliative mode of care on the regular wards, but did not include those formally discharged from the regular wards and readmitted to the hospice ward. In the times studied, there was a decrease in the number of patients who died on the hospice ward, so the trend toward improved mortality could not be easily explained by more frequent diversion of patients to the palliative care service. One limitation of this study is that we were not able to identify a contemporaneous control group for which we had comparable data. The VA uses an electronic medical record and, although we can observe patients in other VA hospitals (e.g., the San Francisco VA), we did not have access to outcome data that we used in this analysis.

In conclusion, an RRS, featuring a physician-led, high-capability medical emergency team, was successfully introduced and sustained in a hospital within the US VA Health Care System. A significant reduction in the rate of cardiac arrest was realized with this intervention, as well as a trend toward lower mortality. Our results suggest that further reductions in morbidity can be realized by use of RRSs throughout the VA network. Larger cooperative studies within the VA system are needed to more definitively assess impacts on mortality. Furthermore, comparison of different team designs among VA centers implementing RRSs may help define certain elements or characteristics of the system that contribute to better acceptance or better outcomes.

Back to Top | Article Outline

AUTHOR CONTRIBUTIONS

GKL helped with study design, conduct of the study, data analysis, and manuscript preparation; LMP helped with statistical analysis; LR helped with data collection and analysis; and THW helped with statistical analysis. All authors attest to the validity and accuracy of the data in this study and approved the final manuscript.

Back to Top | Article Outline

REFERENCES

1. Bellomo R, Goldsmith D, Uchino S, Buckmaster J, Hart G, Opdam H, Silvester W, Doolan L, Gutteridge G. Prospective controlled trial of effect of medical emergency team on postoperative morbidity and mortality rates. Crit Care Med 2004;32:916–21
2. Bellomo R, Goldsmith D, Uchino S, Buckmaster J, Hart GK, Opdam H, Silvester W, Doolan L, Gutteridge G. A prospective before-and-after trial of a medical emergency team. Med J Aust 2003;179:283–7
3. Brilli RJ, Gibson R, Luria JW, Wheeler TA, Shaw J, Linam M, Kheir J, McLain P, Lingsch T, Hall-Haering A, McBride M. Implementation of a medical emergency team in a large pediatric teaching hospital prevents respiratory and cardiopulmonary arrests outside the intensive care unit. Pediatr Crit Care Med 2007;8:236–46
4. Bristow PJ, Hillman KM, Chey T, Daffurn K, Jacques TC, Norman SL, Bishop GF, Simmons EG. Rates of in-hospital arrests, deaths and intensive care admissions: the effect of a medical emergency team. Med J Aust 2000;173:236–40
5. Buist MD, Moore GE, Bernard SA, Waxman BP, Anderson JN, Nguyen TV. Effects of a medical emergency team on reduction of incidence of and mortality from unexpected cardiac arrests in hospital: preliminary study. BMJ 2002;324:387–90
6. Dacey MJ, Mirza ER, Wilcox V, Doherty M, Mello J, Boyer A, Gates J, Brothers T, Baute R. The effect of a rapid response team on major clinical outcome measures in a community hospital. Crit Care Med 2007;35:2076–82
7. DeVita MA, Braithwaite RS, Mahidhara R, Stuart S, Foraida M, Simmons RL. Use of medical emergency team responses to reduce hospital cardiopulmonary arrests. Qual Saf Health Care 2004;13:251–4
8. Galhotra S, DeVita MA, Simmons RL, Schmid A. Impact of patient monitoring on the diurnal pattern of medical emergency team activation. Crit Care Med 2006;34:1700–6
9. Jones D, Bates S, Warrillow S, Opdam H, Goldsmith D, Gutteridge G, Bellomo R. Circadian pattern of activation of the medical emergency team in a teaching hospital. Crit Care 2005;9:R303–6
10. Jones D, Egi M, Bellomo R, Goldsmith D. Effect of the medical emergency team on long-term mortality following major surgery. Crit Care 2007;11:R12
11. Offner PJ, Heit J, Roberts R. Implementation of a rapid response team decreases cardiac arrest outside of the intensive care unit. J Trauma 2007;62:1223–7
12. Sebat F, Johnson D, Musthafa AA, Watnik M, Moore S, Henry K, Saari M. A multidisciplinary community hospital program for early and rapid resuscitation of shock in nontrauma patients. Chest 2005;127:1729–43
13. Sebat F, Musthafa AA, Johnson D, Kramer AA, Shoffner D, Eliason M, Henry K, Spurlock B. Effect of a rapid response system for patients in shock on time to treatment and mortality during 5 years. Crit Care Med 2007;35:2568–75
14. Sharek PJ, Parast LM, Leong K, Coombs J, Earnest K, Sullivan J, Frankel LR, Roth SJ. Effect of a rapid response team on hospital-wide mortality and code rates outside the ICU in a Children's Hospital. JAMA 2007;298:2267–74
15. Ranji SR, Auerbach AD, Hurd CJ, O'Rourke K, Shojania KG. Effects of rapid response systems on clinical outcomes: systematic review and meta-analysis. J Hosp Med 2007;2:422–32
16. Winters BD, Pham J, Pronovost PJ. Rapid response teams: walk, don't run. JAMA 2006;296:1645–7
17. Winters BD, Pham JC, Hunt EA, Guallar E, Berenholtz S, Pronovost PJ. Rapid response systems: a systematic review. Crit Care Med 2007;35:1238–43
18. Wilson NJ, Kizer KW. The VA health care system: an unrecognized national safety net. Health Aff (Millwood) 1997;16:200–4
19. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–9
20. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, Saunders LD, Beck CA, Feasby TE, Ghali WA. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130–9
21. Hillman K, Chen J, Cretikos M, Bellomo R, Brown D, Doig G, Finfer S, Flabouris A. Introduction of the medical emergency team (MET) system: a cluster-randomised controlled trial. Lancet 2005;365:2091–7
22. Chan PS, Khalid A, Longmore LS, Berg RA, Kosiborod M, Spertus JA. Hospital-wide code rates and mortality before and after implementation of a rapid response team. JAMA 2008;300:2506–13
23. Goldhill DR, Worthington L, Mulcahy A, Tarling M, Sumner A. The patient-at-risk team: identifying and managing seriously ill ward patients. Anaesthesia 1999;54:853–60
24. Kenward G, Castle N, Hodgetts T, Shaikh L. Evaluation of a medical emergency team one year after implementation. Resuscitation 2004;61:257–63
25. Parr MJ, Hadfield JH, Flabouris A, Bishop G, Hillman K. The Medical Emergency Team: 12 month analysis of reasons for activation, immediate outcome and not-for-resuscitation orders. Resuscitation 2001;50:39–44
26. Salamonson Y, Kariyawasam A, van Heere B, O'Connor C. The evolutionary process of medical emergency team (MET) implementation: reduction in unanticipated ICU transfers. Resuscitation 2001;49:135–41
27. Jones D, Opdam H, Egi M, Goldsmith D, Bates S, Gutteridge G, Kattula A, Bellomo R. Long-term effect of a medical emergency team on mortality in a teaching hospital. Resuscitation 2007;74:235–41
28. Chan PS, Krumholz HM, Nichol G, Nallamothu BK. Delayed time to defibrillation after in-hospital cardiac arrest. N Engl J Med 2008;358:9–17
29. Chan PS, Nichol G, Krumholz HM, Spertus JA, Nallamothu BK. Hospital variation in time to defibrillation after in-hospital cardiac arrest. Arch Intern Med 2009;169:1265–73
30. Peberdy MA, Ornato JP, Larkin GL, Braithwaite RS, Kashner TM, Carey SM, Meaney PA, Cen L, Nadkarni VM, Praestgaard AH, Berg RA. Survival from in-hospital cardiac arrest during nights and weekends. JAMA 2008;299:785–92
31. Manfredini R, La Cecilia O, Boari B, Steliu J, Michelinidagger V, Carlidagger P, Zanotti C, Bigoni M, Gallerani M. Circadian pattern of emergency calls: implications for ED organization. Am J Emerg Med 2002;20:282–6
32. DeVita MA, Bellomo R. The case of rapid response systems: are randomized clinical trials the right methodology to evaluate systems of care? Crit Care Med 2007;35:1413–4
33. Devita MA, Bellomo R, Hillman K, Kellum J, Rotondi A, Teres D, Auerbach A, Chen WJ, Duncan K, Kenward G, Bell M, Buist M, Chen J, Bion J, Kirby A, Lighthall G, Ovreveit J, Braithwaite RS, Gosbee J, Milbrandt E, Peberdy M, Savitz L, Young L, Galhotra S. Findings of the first consensus conference on medical emergency teams. Crit Care Med 2006;34:2463–78
34. Downey AW, Quach JL, Haase M, Haase-Fielitz A, Jones D, Bellomo R. Characteristics and outcomes of patients receiving a medical emergency team review for acute change in conscious state or arrhythmias. Crit Care Med 2008;36:477–81
35. DeVita MA, Schaefer J, Lutz J, Dongilli T, Wang H. Improving medical crisis team performance. Crit Care Med 2004;32:S61–5
36. Peberdy MA, Kaye W, Ornato JP, Larkin GL, Nadkarni V, Mancini ME, Berg RA, Nichol G, Lane-Trultt T. Cardiopulmonary resuscitation of adults in the hospital: a report of 14720 cardiac arrests from the National Registry of Cardiopulmonary Resuscitation. Resuscitation 2003;58:297–308
© 2010 International Anesthesia Research Society