Rapid Response Teams Do They Make a Difference?

Research DIMENSION

Jenny Jolley, MSN, RN, CCRN; Heather Bendyk, BS; Bonnie Holaday, DNS, RN, FAAN; Kristine A. K. Lombardozzi, MD, FACS; Corinne Harmon, MS, EdD, RN, AOCN

Hundreds of lives are now being saved in hospitals across the country with the use of rapid response teams. These teams are composed of clinicians who bring critical care expertise to the patient bedside. The purpose of these teams is to assess and stabilize the patient, assist with communication among the interdisciplinary care providers, educate and support the staff caring for the patient, and assist with transfer of the patient if necessary. Research has shown that, with successful implementation of a rapid response team, the percent of codes and mortality rates decrease. The purpose of this study was to evaluate the effectiveness of implementing a rapid response team at 1 medical center. The results from the study demonstrated a decrease in the percent of codes outside the critical care units. However, it did not show a decrease in overall mortality rates for the patients. Data review will continue as we strive to improve our overall mortality rates while maintaining a decrease in the amount of codes. Keywords: Rapid response teams, Cardiopulmonary arrest, Emergency care

[DIMENS CRIT CARE NURS. 2007;26(6):253/260]

When someone calls out, ‘‘Call a code!’’ everyone hospital mortality. The goal of the IHI is to reduce in- knows there is a patient in crisis. Code team members patient deaths by 100,000 a year. This number is based work in unison to save these patients. Now there is a on the number of hospitals participating in the 100K new concept that is gaining momentum that can make Lives Campaign. The IHI states that 75 lives a year crash carts obsolete. Rapid response teams (RRTs) or could easily be saved at a moderate-size hospital with emergency medical teams focus on preventing a patient 15,000 patient admissions annually. Forty-five of those crisis by addressing changes in patient status before a lives could be saved with the use of an RRT.1 cardiopulmonary arrest occurs. An RRT is a team of clinicians who bring critical care expertise to the patient’s bedside. Daily people die The goal of the Institute for unnecessarily in our hospitals. The implementation of HealthCare improvement is to an RRT is one of the 6 recommended healthcare reduce inpatient deaths by initiatives of the Institute for HealthCare Improvement 100,000 a year. (IHI), which can be found on their Web site at IHI.org. The use of these teams has shown to dramatically alter

November/December 2007 253

The concept of RRTs started as an experiment intro- METHODOLOGY duced in 1995 by an Australian hospital and quickly spread worldwide because of its proactive approach to Research Questions and Hypotheses increasing patient safety. This same hospital conducted Based on the review of the literature, the following re- a study using a cluster randomized controlled trial search questions were asked: (1) Do RRTs make a dif- that consisted of 23 hospitals, 11 were control and ference in the number of codes called outside the critical 2 12 were medical emergency teams. care units; (2) do RRTs make a difference in mortality The findings from this study indicated that an RRT rates on inpatient units outside critical care? greatly increased the calls to the emergency team but did An RRT is defined as a critical care nurse and a not substantially affect the incidence of cardiac arrest respiratory therapist called to the floor based on call- or unplanned intensive care unit (ICU) admissions of ing criteria for activation to assist with a change in a 2 unexpected deaths. This study examined the effects of patient’s condition outside the critical care units. A code the introduction of an RRT and concluded that this did is defined as when a healthcare member identifies a not greatly improve their study outcomes. Therefore, patient or individual (ie, visitor and employee) as having based on these outcomes, no process change was imple- a cardiopulmonary collapse. Mortality rates are deter- 2 mented at these hospitals. Some possible explana- mined based on the ratio of deaths to total population. tions for their findings included (1) wrong outcomes Two hypotheses were tested: (1) There will be no studied, (2) inadequate implementation of the RRT statistically significant difference in the percent of codes system, (3) the possibility that the control hospital’s called on hospital inpatient units after the implemen- outcomes were influenced as a result of being in the tation of an RRT than the percent of codes called on study, (4) the hospitals studied were unrepresentative, or inpatient units before the implementation of the RRT; (5) the study did not have adequate statistical power and (2) there will be no statistically significant difference 2 to detect important treatment effects. Even though the in the mortality rates on hospital inpatient units after study findings were not significant, it did lead other the implementation of an RRT than there were before hospitals to explore this option. the implementation of the RRT. Later studies of RRTs found that they are associ- The following demographic variables were also evaluated with a 50% reduction in non-ICU arrests, reduced ated: Patient’s age, length of stay, sex, ethnicity, admit- postoperative emergency ICU transfers by 58%, and re- ting source, and severity of illness. These variables were 3 duced postoperative emergency ICU deaths by 37%. evaluated to determine if there was any significant dif- Some institutions have shown decreased mortality and ference between the groups of patients in 2005 and 2006. cardiac arrest in hospitalized patients with an accom- panying lowered relative risk for respiratory failure (79%), for stroke (78%), for severe sepsis (74%), and Research Design for acute renal failure (88%) with the implementation This study used a quasi-experimental design. There was of an RRT. Others have shown reductions up to 30% neither randomization nor a control group. The study and 27%, respectively, for cardiac arrests and unex- used a time series design with preexperimental observa- pected deaths.4,5 tions before the implementation of the RRT and Research has shown that almost all critical inpatient postobservations after the implementation of the RRT. events are preceded by warning signs for an average of The sample consisted of all inpatient units and excluded 6 to 8 hours before arrest.6 Such warning signs in a all critical care units at a regional medical center. The patient’s condition include, but are not limited to the data were collected for 12 months before the implemen- following: tation of the RRT and for 12 months after the implementation of the RRT. & Change in vital signsVa low or elevated heart rate, blood pressure, or respiratory rate Study Facility & New onset of difficulty breathing Spartanburg Regional Healthcare System is a 488- & Change in level of consciousness licensed bed, not-for-profit teaching medical center & Repeated or prolonged seizures located in upstate South Carolina. The average daily & Low urinary output adult census is 390, with a total discharge rate averaging & Acute decrease in oxygen saturation despite delivery approximately 2,400 patients per month. Patients are of oxygen admitted to the hospital from the emergency room, phy- Any of these changes are criteria for calling the sician offices, and transfers from other hospitals. This RRT.7 Regional Emergency Center is one of the busiest Level I

254 Dimensions of Critical Care Nursing Vol. 26 / No. 6

Trauma Centers in the nation with more than 100,000 care nurse and a respiratory therapist with their primary patients in the past year. goal being to support and to assist the staff on the floors with any concerns they have regarding changes in their COMPOSITION AND PURPOSE OF THE TEAM patient’s condition. Spartanburg Regional made the decision to join IHI’s initiative of 100K Lives Campaign. In August of 2005, an 8-member planning committee consisting of a nurse The participants included all manager, a clinical unit educator, risk management, res- patients admitted to any inpatient piratory therapists, and the nursing director of surgical unit within the hospital. services was convened to begin examining the concept of an RRT. We needed the expertise from all members to brainstorm, benchmark, and develop the best way to initiate and develop an RRT for the hospital. During The planning committee for the RRT developed a this time, we discovered that we already had an informal set of ‘‘calling criteria’’ for the activation of the team. RRT in place. The respiratory therapists stated that they This included the recommended reasons for activation received calls from the medical surgical floors on a daily along with telephone numbers for the team members, basis to assist staff in potential code situations. The listed on a pocket card (see Figure 1). On the back of the hospital administrators were there to lend support and pocket card is a physician communication tool designed provide any additional guidance and ensure successful to help staff communicate effectively with the physician. implementation of the team. The type of communication system used is called SBAR. After numerous weekly and monthly meetings, we The S stands for situation, B stands for background, ‘‘went live’’ with our RRT on January 2, 2006. The par- A stands for assessment, and R stands for recommen- ticipants included all patients admitted to any inpatient dation (see Figure 1). SBAR communication helps staff unit within the hospital. The team consists of a critical give a more concise and complete report to the physician.

Figure 1. Pocket card developed by Planning Committee at Spartanburg Regional Medical Center.

November/December 2007 255

Any staff member in the hospital has the capability to tion. If necessary, the team is able to assist with transfer activate the RRT by calling the designated numbers. of the patient to the ICU. According to the literature, 3 problems often lead to failure to rescue a patient in a timely manner. These Data Collection Procedures 3 problems are as follows: (1) failure to recognize a The Institutional Review Board indicated that this study problem; (2) failure to appropriately communicate the did not require their review. The data were collected problem whether it is patient to staff, staff to staff, or from the patient’s chart each time the RRT was acti- staff to physician; and (3) failure in planning and the vated. The planning committee developed a data collec- ability to treat the problem, which includes assessment, tion tool (see Figure 2). This tool is completed by the treatments, and establishing goals.4 staff nurse and submitted for review. This tool is in Upon activation, the RRT immediately responds to triplicate form with each page indicating where the the call to assist staff on the floors by bringing their additional copies are to be sent for review. The tool is expertise to the bedside. By doing so, the RRT is able to verified and checked for completeness by an assigned assist staff in providing the appropriate care for the auditor. Upon review of the data, if the auditor needs to patient who has demonstrated changes in their condi- follow up with any of the involved units for clarification,

Figure 2. Data collection tool developed by Planning Committee at Spartanburg Regional Medical Center.

256 Dimensions of Critical Care Nursing Vol. 26 / No. 6

Copyright @ Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Rapid Response Team he or she will contact the appropriate person for clarification. Then he or she enters the data into an electronic database system used by the hospital. Monthly, a report is prepared by the Quality Assur- ance Department with data entered previously. These data are presented bimonthly to the Resuscitation Com- mittee, the Critical Care Committee, and to the hospital administration to keep them informed on the progress the RRT is making. These were the data set used for the entire study. Graph 2. Percent of codes outside critical care units.

Results and 3). Therefore, the null hypothesis of no difference The SAS Jmp software was used for data analysis. Data was not supported. were reviewed from January 2006 to January 2007 from However, the data for the attribute variables (see all in patient units outside critical care. The results indi- Table 1) did not find any statistical difference. As cated that 76 calls (see Graph 1) were placed to activate demonstrated in Table 1, there is no significant differ- the RRT. The reports from staff cited acute changes in the ence in the mean age nor any significant difference patient’s condition. The changes the staff reported were: between length of stay, sex, ethnicity, admitting source, or all patient refined diagnostic related groups severity Reported Change Percent Reported of illness. The severity of illness and the P value of .1317 also indicated no significant difference in the se- Mental status 26% verity of illness in the 2 patient populations under study Respiratory rate 20% (see Table 1). The severity of illness is determined by Blood pressure 16% physician diagnoses using the 3M All Patient Refined Staff concerned/worried 13% Diagnostic Related Groups system. Thus, results did not indicate a difference in the times the RRT was activated Heart rate 12% by the attribute variables. G Sustained Sp02 85% 8% The results from the activation of the RRT found Acute significant bleeding 3% that 57% of the patients were transferred to the critical Failure to respond to treatment 2%

The data from the dependent variables (percent of codes called outside the critical care units) demonstrated a statistical difference in the percent of codes called (P = .0262) outside the critical care units between 2005 and 2006. In 2005, the mean number of codes was 66.78% (n = 161), with a median of 69.91. In 2006, the mean was 51.37% (n = 139), with a median of 48.69%. This was a 21% decrease from 2005 in the percent of codes called outside the critical care units (see Graphs 2

Graph 1. Number of calls placed to active RRT. Graph 3. Percent of codes outside critical care units.

November/December 2007 257

the hospital, protocols used, and the type of study being TABLE 1 Percentage Analysis and 7-14 Demographic Data performed. For example, one group of investigators at another large medical center performed a study that CY05 CY06 P demonstrated an overall decrease of 14.95% in dis- Percent of codes outside critical care Units charge rates for coded patients with overall mortality rates dropping by 32%.9,15 All the institutions, with the Analysis of % of codes outside critical .6678 .51372 .0262* exception of one, reported positive outcomes with the care units successful implementation of an RRT. These organiza- Demographics tions noted decreases in their overall cardiac arrest rates Age, mean 60.92 63.89 .1758 and overall mortality rates, with some of the institutions Length of stay, mean 7.81 8.80 .4959 demonstrating a 50% reduction in non-ICU arrests along with a reduction in arrest before an ICU transfer.16 Sex The purpose of the study at this medical center was Male 93 76 .6410 to determine the effectiveness of implementing an RRT. Female 68 63 The stated hypothesis that there would be no difference Ethnicity in the percent of codes called outside the critical care units in 2005 when there was no formal RRT and in African American 44 33 .6873 2006 when a formal RRT was functioning was not sup- White 115 105 ported. However, the null hypothesis regarding no dif- Hispanic 2 1 ference in mortality rates was supported. Admitting source Study Limitations MD referral 30 22 .7940 We identified several study limitations to the study. Transfer from another hospital 8 8 First, there was difficulty obtaining physician support. EC 122 108 Some physicians thought that an RRT would interfere APR DRG severity of illness with their care for their patient. The physicians did not have a realistic understanding of the role of the team. 1 3 2 .1317 The team’s role to support staff and assist with the care 288of the patient based on orders received from the attend- 34223ing physician was not understood by all physicians; thus, 45260the idea of an RRT was not accepted by all physicians. The protocols for the RRT were nurse driven, and EC indicates emergency center; APR DRG, all patient refined diagnostic related we did not have a physician on the team. Therefore, groups. *Statistically significant at " = .05 level. the functions of the critical care nurse and respiratory therapist were limited in what they could do for the care units and 35% remained on the floor. Only 1% patient outside the critical care units by hospital policy had resulted in a code called. There was no supporting and their scope of practice. documentation available on the other 7% of patients Another area identified as a study limitation was requiring assistance from the RRT. Some of the recom- that our ‘‘go-live’’ was limited to inpatients only. We mendations or interventions used by the RRT during also identified that a nonstandardized version of the these calls included insertion of oral airways, drawing RRT was in use before actual implementation of an arterial blood gases, fluid boluses, initiation of vaso- identified RRT. Thus, it took some time for the nursing active drips, administration of diuretics, and sedation. staff to switch to the formal structure that included a Mortality rates at the Regional Medical Center are nurse instead of just a respiratory therapist. based on ratio of deaths per 1,000 discharges. In 2005, the Finally, this is one hospital in one geographic loca- mortality rate was 2.92%, and in 2006, the mortality rate tion. We feel that this study needs to be replicated in was 2.93% Therefore, the null hypothesis was supported. multiple hospitals with experimental controls in place. Some identified biases in the study that could affect Discussion the results included increased nursing care hours plus The review of the literature on RRTs on overall survival all nursing units were self-governing meaning that they rates in hospitals showed significant decreases in mor- must provide their own staff to cover their units. All tality rates. These results varied from institution to nursing units were using newly purchased defibrillators institution based on the number of admissions, size of with automated external defibrillators. Upon receipt of

258 Dimensions of Critical Care Nursing Vol. 26 / No. 6

Copyright @ Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Rapid Response Team the new equipment, there was extensive education on CONCLUSION codes and code documentation. This education could In summary, the review of the literature suggested differ- have highlighted patient’s code status and could have ent utilizations for an RRT. Each organization repre- increased staff perception regarding patients who are at sented, with one exception, found positive outcomes risk for arrest. with the successful implementation of an emergency medical team. These organizations reflected a difference IMPLICATIONS FOR CLINICANS in their overall cardiac arrest rates and overall mortality Implementation of such a team at any organization can rates. Some organizations experienced dramatic decreases. help build trusting relationships between nurses and In the hospital where the RRT was not effective, there ancillary staff along with empowering these staff seemed to be a failure of inadequate implementation. members.7 One of the goals for the registered nurse on Therefore, based on this research review regarding what the RRT is not to take over the role of the bedside nurse is the impact of implementing an RRT, the hypothesis is but to keep him or her involved and to use this oppor- that there is a difference in the number of cardiopulmo- tunity to teach and mentor more advanced assess- nary arrest and overall mortality rates. ment and intervention skills. This leads to a tremendous Spartanburg Regional demonstrated a decrease in growth in assessment skills of the nursing staff on the the percent of codes outside the critical care units. medical and surgical units.4 Another benefit of this form However, it did not show a decrease in overall mortality of mentoring is to sharpen the critical thinking skills rates for our patients. The benefits achieved with the of staff nurses and to help both nurses and the units gain RRT were improved patient safety, fewer cardiopulmo- confidence in interdisciplinary collaboration.8 nary arrests, and an increased awareness among staff Mutual respect between RRT members and staff members to identify and report changes in a patient’s nurses on the floor is also gained along with the elimi- condition. This study found that activating an RRT nation of invisible barriers among nursing staff that makes a difference. As demonstrated in our study, an allowed working relationships and trust to be built.4 RRT played an important role in preventing cardiopul- Staff nurses discovered that they will not be criticized monary arrests. for calling for help and that they have also learned to be In summary, the development and implementation more sensitive to opportunities to rescue patients before of RRTs is a concept that is gaining acceptance. These they get into serious trouble.15 teams can promote a culture of safety by building team- Another positive outcome of the use of RRT is that work and spreading knowledge and skills throughout it provides immediate clinical consultation to all nurses hospitals. The goal of all these teams is to improve in every unit when they have an urgent situation that patient outcomes and decrease mortality. Rapid response needs intervention.7 The nursing staff feels empowered teams help to create healthcare systems of the future. to call and activate the RRT to provide additional sup- Studies have shown that hospital mortality rates decrease port and backup as needed for their patients. with the use of an RRT.

References 1. Comarow A. Saving lives hospitals have signed on to a six-part Another positive outcome of the plan to avoid a multitude of unnecessary deaths. US News World Rep. 2005:74-81. use of RRT is that it provides 2. Hillman K, Chen J, Cretikos M, et al. Introduction of the immediate clinical consultation medical emergency team (MET) system: a cluster-randomised controlled trial. Lancet. 2005;365:2091-2097. to all nurses on every unit when 3. Louisiana Hospital Association. Preventing avoidable deaths with rapid response teams. 2005. Available at: http://www. they have an urgent situation lhaonline.org. Accessed August 2005. that needs intervention. 4. Simmons JC. Buying time, saving lives: hospital rapid response teams find ways to reduce mortality outside the ICU. The Quality Letter for Healthcare Leaders. Lippincott, Williams, and Wilkins; 2004. 5. Scholle C, Mininni N. Best-practice interventions: how a rapid Nursing leadership at any hospital must determine response team saves lives. Nursing. 2006;36:37-40. 6. Clinical Decision Support Spotlight. Implementing rapid the most difficult obstacles to overcome and then iden- response teams: the early experience of two MIDAS+ hospitals. tify how to overcome these obstacles, as did the leader- 2005;(3):1-7. ship of Spartanburg Regional. Without the support of 7. Kleinpell R. Redefining resuscitation [Electronic version]. Nursing Spectrum. Career Management Magazine; 2005. nursing leadership and hospital administration, this type 8. Ball C, Kirkby M, Williams S. Effect of the critical care of team cannot be successful. outreach team on patient survival to discharge from hospital

November/December 2007 259

and readmission to critical care: non-randomized population ABOUT THE AUTHORS based study [Electronic version]. BMJ. 2003;327:1014. Jenny Jolley, MSN, RN, CCRN, is a recent graduate of the MSN 9. Bellomo R, Goldsmith D, Shigehiko U, et al. A prospective before-and-after trial of a medical emergency team [Electronic program at Clemson University. She has 14 years of critical care version]. Med J Aust. 2003;179:283-287. experience and is employed with Spartanburg Regional Medical 10. Bunch D. ‘‘RRTs’’ to the rescue! AARC Times. 2005:30-35. Center in South Carolina. 11. DeVita M, Braithwaite R, Mahidhara R, Stuart S, Foraida M, Heather Bendyk, BS, is a statistician at Spartanburg Regional Simmons R. Use of medical emergency team responses to reduce hospital cardiopulmonary arrests [Electronic version]. Medical Center and assists with the data collection process and Qual Staff Health Care Online. 2004;13:251-254. reporting of data involving the rapid response team. 12. Hellwig S, Piper L, Naylor E. Forty hours under pressure: a Bonnie Holaday, DNS, RN, FAAN, is a professor in the School rapid response improvement team achieve synergy. J Healthc of Nursing and Institute on Family and Neighborhood Life at Qual. 2002;24:21-24. 13. Mahn-DiNicolav. Implementing rapid response teams: the early Clemson University. Dr Holaday serves as a consultant with experience of two MIDAS+ hospitals. ACS MIDAS. 2005;3:1-7. DCCN. 14. Pickoff R. Are rapid response teams simply a bandage on a Kristine A. K. Lombardozzi, MD, FACS, is the Director of Surgical bigger problem? Physician Exec. 2006;3:36-38. Critical Care at Spartanburg Regional Medical Center. 15. Baptist Memorial develops medical response team to combat mortality rates. Target Quality. The Maryland Hospital Asso- Corinne Harmon, MS, EdD, RN, AOCN, is an instructor at Clemson ciation Quality Indicator Project; 2005:4-6. University. 16. Buist M, Bernard S, Anderson, MB. Epidemiology and prevention of unexpected in-hospital deaths. 2003. Available Address correspondence and reprint requests to: Jenny Jolley, at: http://www.rcsed.ac.uk/journal/svoll5/10500003.html. MSN, RN, CCRN, 585 Reynolds Road, Chesnee, SC 29323 Accessed October 30, 2005. ([email protected]).

Coming in the January/February 2008 Issue

& Assessment of Attitudes of Intensive Care Unit Staff Toward Clinical Practice Guidelines & Utilizing the Synergy Model as Best Practice in Endotracheal Tube Suctioning of Critically Ill Patients & Healthy Work Environment & News Bits & Research Abstracts & Spirituality: A Dimension of Holistic Critical Care Nursing & Transfusion-Related Acute Lung Injury & Comparing Tympanic Temperatures in Both Ears to Oral Temperature & Florence Nightingale

260 Dimensions of Critical Care Nursing Vol. 26 / No. 6