The Joint Commission Journal on Quality and Patient Safety Care Processes

An Institutionwide Approach to Redesigning Management of Cardiopulmonary Resuscitation Geoffrey K. Lighthall, MD, PhD; Michael Mayette, MD, FRCPC; T. Kyle Harrison, MD

A

n estimated 350,000–750,000 adult, in-hospital cardiac arrest events occur annually in the United States.1 Despite widespread training in basic life support (BLS) and advanced cardiovascular life support (ACLS) among hospital personnel, survival from in-hospital cardiac arrests is approximately 15%,1–5 reflecting, for example, underlying conditions, inadequate monitoring and detection, and the quality of resuscitation itself.6 Factors found to impair the success of in-hospital cardiopulmonary resuscitation (CPR) include inadequate depth, rate, or consistency of chest compressions; overventilation; delayed defibrillation of pulseless ventricular arrhythmias; and delays in the administration of drugs indicated by ACLS algorithms.7–10 To achieve the highest likelihood of success, the emergency must be identified, and immediate life-saving CPR must then be started and the code team activated. In assembling (often asynchronously), the code team must then begin emergent cardiac care. If, at any step in the process, a team member fails to respond appropriately, the overall care of the patient may suffer. In 2005, before the start of our resuscitation redesign project, our hospital, the US Department of Veterans Affairs (VA) Palo Alto (California), a tertiary medical center, had organized a rapid response system, with a medical emergency team (called the “eTeam”) with the goal of using specific warning criteria and nurse and family concern as triggers for summoning the primary and ICU–based teams.11 The eTeam has led to a sustained 50% reduction in the rate of cardiac and respiratory arrests in the hospital since its inception in 2005. The arrest frequency is approximately 20 per year. Yet rather than write off this population as being a very small and extremely sick group with a high frequency of ultimately fatal prognoses, in 2006 we decided to seek out areas where existing programs such as ACLS training, nursing competency fairs, and annual training modules have fallen short of creating true expertise and excellence in care. Because an arrest represents a human’s last chance for survival, and given the availability of specific guidelines based on the consensus of the science of proper management,12,13 we considered it important

Article-at-a-Glance Background: Despite widespread training in basic life support (BLS) and advanced cardiovascular life support (ACLS) among hospital personnel, the likelihood of survival from in-hospital cardiac arrests remains low. In 2006 a universityaffiliated tertiary medical center initiated a cardiopulmonary (CPR) resuscitation redesign project. Redesigning the Hospital’s Resuscitation System: The CPR Committee developed the interventions on the basis of a large-scale view of the process of delivering BLS and ACLS, identification of key decision nodes and actions, and compartmentalization of specific functions. It was proposed that arrest management follow a steady progression in a twolayer scheme from BLS to ACLS. Handouts describing team structure and specific roles were given to all code team providers and house staff at the start of their month-long rotations. To further increase role clarity and team organization, daily morning and evening meetings of the arrest team were instituted. Site-specific BLS training, on-site ACLS refresher training, and defibrillator training were initiated. Project elements also included use of unannounced mock codes to provide system oversight; preparation and distribution of cognitive aids (printed algorithms, dosing guides, and other checklists to ensure compliance with ACLS protocols), identification of patients who may be unstable or a source of concern, event review and analysis of arrests and other critical events, and a CPR website. Conclusion: A mature hospital-based resuscitation system should include definition of arrest trends and resuscitation needs, development of local methods for approaching the arresting patient, an emphasis on prevention, establishment of training programs tailored to meet specific hospital needs, system examination and oversight, and administrative processes that maximize interaction between all components.

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

157

The Joint Commission Journal on Quality and Patient Safety for our hospital to deliver care according to these standards.14 Before we undertook the redesign project, the resuscitation system was a patchwork of technology; training; and documentation, including policies intended to provide up-to-date emergency cardiovascular care. Management of arrests seemed chaotic, and providers often lacked specific instructions on their roles and knowledge of how the whole operation was designed to function. All training in ACLS and BLS was conducted externally and, as is often the case with such generic courses, lacked consideration of true workplace conditions and challenges. Following implementation of the hospital’s eTeam, we began to appreciate what could be achieved by a more coordinated system of care15 and decided to reexamine our approach to resuscitation and whether any opportunities to improve care existed. To determine what potential roadblocks and or hurdles to high performance existed for cardiac arrest teams, we mapped out each step for the successful management of a patient in cardiac arrest and then identified potential barriers as well as strategies for success. We then created policies, procedures, and training programs to facilitate ideal performance in resuscitation management. Innovations in resuscitation have come mainly in the form of studies on single interventions, such as CPR feedback, cardiocerebral resuscitation16,17 and use of capnography,18–20 that improve circulatory function, survival, or other outcomes. Our goal was to not only incorporate all compelling scientific advances into our practices but seek the best means of implementation. Despite a growing interest in defining best practices related to resuscitation, we found that there is currently little scientific evidence on which to draw. In this article, we describe a series of improvements to our resuscitation system based on prospective evaluation of potential failure points and the institution of remedies through new policies and training programs and other elements. The redesign evolved in an ongoing effort to improve care delivery. We do not presume to present our approach as definitive but rather as one that other heath care organizations may learn from as they assess the need for improvement of their own resuscitation systems.

Redesigning the Hospital’s Resuscitation System SETTING The VA Palo Alto is a tertiary medical center serving a geographical area of 13,500 square miles. The 150 acute care beds include 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 4,500 acute medical/surgical admissions to the hospital each year 158

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 and overseen by physicians board certified in intensive care.

STARTING THE INITIATIVE As the leaders of the hospital’s CPR Committee, the authors began the process in April 2006 by taking a broad view of in-hospital resuscitation, which entailed consideration of the needs of a patient arresting on the ward and continuing through resolution of the arrest—as represented in the flowchart shown in Figure 1 (page 159). We posed open-ended questions, such as What is the most likely type of arrest? What should the first responder do? What are the key decision points? What are the key skills needed for managing an arrest; and who should be trained for each? We then asked whether our current organizational and training methods could guarantee excellence in the areas described. We generated the list of needs and subprograms outlined in Table 1 (page 160), which we expand on in this article.

THE TWO-LAYER CONCEPT In published studies, the majority of in-hospital cardiac arrests have pulseless electrical activity, aystole, or symptomatic bradycardia at the time of discovery.4 Examination of arrests in our facility revealed a similar pattern (Table 2, page 161). Therapy for nonshockable cardiopulmonary arrests involves delivery of effective CPR while commencing treatment for a host of underlying abnormalities. The number of tasks that need to be accomplished can be overwhelming for any health care worker, particularly those with only sporadic exposure to critical incidents. To ensure that arresting patients receive the proper therapy in a reproducible manner, we proposed that arrest management follow a steady progression from BLS to ACLS. With this twolayer scheme, we emphasize that it should be the role and training goal of floor nurses to practice and master the immediate delivery of chest compressions, and later, ventilation as additional colleagues and equipment become available. The subsequent ACLS layer is provided a few minutes later by the arriving code team and ensures continuation of the proper compression/ventilation sequence while adding full functionality to the resuscitation. The arrest team includes specific respondents who perform chest compressions (interns), freeing the

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

The Joint Commission Journal on Quality and Patient Safety Flowchart Describing Course of a Deteriorating Patient, with System Redesign Components

Figure 1. System redesign components relating to specific care needs and intervals are indicated on the margins. The basic life support (BLS) layer consists of the first responders’ arrival to an arrest and their provision of chest compressions while additional equipment and personnel are assembled. The advanced cardiac life support (ACLS) layer evolves as the code team arrives and organizes itself into airway management, cardiopulmonary resuscitation (CPR), therapy, and leadership subdomains. Italics denote specific training programs or activities directed to various components and subcomponents of emergency care. PEA, pulseless electrical activity; RT, respiratory therapist.

initial nurse respondents to gather and present pertinent information. BLS Phase. Discovery of a “down” patient should lead to an immediate pulse check. In the absence of a pulse, the first responder should immediately call for help and begin chest compressions. The second responder should ensure that the code blue system was activated and then provide bag-mask ventilation and assist with compressions if needed. The broader call for help should quickly lead to the delivery of a crash cart and for other responders to obtain and use the backboard, turn on the defibrillator and attach the pads, and provide chest compressions as needed. We based these actions on three sets of findings. First, a wealth of anecdotal reports at our hospital indicated that many non-ICU nurses, even if ACLS–trained, did not execute the ap-

propriate action when faced with a pulseless patient—and even after all medical/surgical ward nurses were trained in CPR. Insistence on performance of CPR to the exclusion of ACLS seemed to be an effective approach to eliminate the emotion, stress, and uncertainty that often lead to inactivity and to ensure initiation of therapy beneficial to an arrest victim.21,22 Perhaps the expectation that nurses who do not work with intravenous (IV) cardiac medications and analyze cardiac rhythms on a daily basis learn abnormal cardiac rhythms and accompanying drug/shock sequences had a diminished return on investment for the training. Second, even with an arrythmia amenable to defibrillation (that is, a shockable rhythm), it takes time to obtain the crash cart and evaluate the presenting rhythm. Effective CPR during this time improves oxygen delivery and adenosine triphosphate generation in myocardial tissue, which improves the efficacy of defibrillation, so it should not be thought of as a second priority in an unmonitored arrest.23–25 Reluctance of ACLS–trained nurses to independently defibrillate may be widespread.26 Finally, ACLS guidelines released in 2005 emphasized CPR as the most important intervention in nonmonitored arrests. Data from our hospital showed that arrests presenting with shockable rhythms represented only 10%–15% of the total number of patients, thereby lending support to the idea of making CPR rather than defibrillation an initial priority. This formulation of floor-arrest priorities created some controversy at our hospital at the time. The 2005 ACLS guidelines did not specifically address care of the nonwitnessed/nonmonitored in-hospital arrest victim27,28; however, the 2010 guidelines did, essentially validating our approach.14,29 ACLS Phase. A code call will summon a number of personnel with essential roles in managing the complete set of ACLS algorithms. The transition from a CPR–only or CPR/ventilation scheme to a full-team effort needs to be made quickly and seamlessly as new personnel are needed (1) to relieve the persons performing CPR and (2) to deliver more complex therapy that includes IV medications, advanced airway devices, and defibrillation. The code team, after it is assembled, then identifies the correct ACLS treatment algorithm and continues the management in accordance with those guidelines. We emphasize the use of laminated cards with algorithms and expect the team to use such cognitive aids throughout the resuscitation. The 2010 set of algorithms was reduced to two key pathways—(1) a shockable pathway that emphasizes defibrillation and (2) a nonshockable pathway based on establishing sinus rhythm and searching for reversible causes of pulseless electrical activity.14,30 Before the start of our redesign efforts, expectations of person-

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

159

The Joint Commission Journal on Quality and Patient Safety Table 1. Main Elements and Interventions Instituted in the Cardiopulmonary Redesign Program* Main Element Training that reflects the epidemiology and outcome of arrests Two-layer concept of arrest management Recognition of roles and responsibilities

Intervention Data aggregation and analysis, periodic review of literature, development of training programs by the CPR Committee BLS and ACLS refresher training Team meetings Handouts for rotating residents Mock codes Simulation training in BLS, ACLS BLS and ACLS refresher training Creation of defibrillator course Team meetings Preparation of cognitive aids Mastery training on defibrillator Posters in work areas Demo crash cart Training with actual equipment Development of hospital-specific cognitive aids Review of literature; automated literature searches Mock codes, debriefings CPR website Expansion of number of RNs providing ACLS training On-site ACLS training Integration of CPR and medical emergency team administration and data analysis

Improvement of conduct of arrests

Improve familiarity of equipment

Improvement of compliance with ACLS algorithms Integration of new ideas and technology System integration and oversight

Arrest prevention

* CPR, cardiopulmonary resuscitation; BLS, basic life support; ACLS, advanced cardiovascular life support.

nel at critical junctures of arrest care were documented in the hospital’s CPR policy, but there was little effort to reinforce and rehearse the specific tasks. Most resident trainees feel underprepared to lead arrest teams.31 ICU attendings and fellows did not carry pagers and attend arrests and often went to codes without a defined role, which led to confusion in the team, particularly with regard to leadership. We now reiterate team design and specific BLS and ACLS roles at team meetings, in training exercises, and on posters (placed in hallways and break rooms). Code blue team members and their specific roles are shown in Table 3 (page 161). Although BLS and ACLS might together imply a two-layer approach, as far as we can tell, our implementation of such an approach may be unique in terms of the delineation of specific CPR administration responsibilities at different phases of the arrest on the basis of personnel and situational variables.

ROLE RECOGNITION AND TEAM MEETINGS Cardiac arrest teams, particularly house-staff arrest teams, often face the challenge of establishing leadership and a team structure amid the chaos of members’ arrival from all directions. Arrest simulations have shown more timely attainment of important milestones such as first shocks, uninterrupted CPR, and drug administration when the arrest team is allowed to organize 160

briefly before getting to work.9,32–34 Our own studies of in situ arrest simulation have supported this lack of leadership and role identification.35 To address this issue, in July 2009 we produced printed handouts describing team structure and specific roles, which were given to all code team providers and all house staff at the start of their month-long rotations. In mid-2011, having recognized that this intervention failed to increase role clarity and organization at codes, we instituted daily 10-minute 8:00 A.M. morning meetings of the arrest team. Members are summoned by code pager to a convenient location, where the following activities ensue: ■ Members introduce themselves and state their role on the team. ■ A quick check is made to see who has a copy of the in-house ACLS algorithms; use of such cognitive aids is reinforced. ■ A brief teaching point is made by either the resident leader, fellow, or pharmacist. ■ To emphasize prevention, an inquiry is made about whether any member has heard of any unstable patients who would need to be seen after the meeting. Four to five months after the institution of these meetings, they seemed to be generally accepted and led to a request to also conduct evening meetings for nighttime and on-call staff. The 10:00 P.M. meeting follows a 9:00 P.M.–10:00 P.M. handoff

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

The Joint Commission Journal on Quality and Patient Safety Table 2. Epidemiology and Outcomes for Non-ICU Arrests*

Table 3. Code Blue Team Members, Daytime*

Metrics and Demographics (October 2004–September 2011) Demographics Acute care beds/ICU/stepdown beds (no.) Age (mean, ± SD) Sex (M/F) Medical unit Surgical unit Other unit/emergency room Initial Rhythm Pulseless ventricular tachycardia Ventricular fibrillation Pulseless electrical activity Asystole Respiratory arrest Time of Day 0:00–7:59 (12:00–7:59 A.M.) 8:00–15:59 (8:00 A.M.–3:59 P.M.) 16:00–23:59 (4.00 P.M.–11:59 P.M.) Time of Week Weekday Weekend Witnessed and/or Monitored Witnessed and monitored Witnessed, not monitored Not witnessed, monitored Not witnessed, not monitored

150/15/15 68.0 ± 12.1 99.2%/0.8% 50.8% 21.8% 27.4% 7.3% 9.7% 54.0% 14.5% 14.5% 33.9% 46.0% 20.2% 71.4% 28.6% 46.8% 26.6% 4.8% 21.8%

Outcomes Before (October 2004–September 2006) and After (October 2006–September 2011) Cardiopulmonary Resuscitation Redesign

Arrests per year (mean no.) ROSC, ICU transfer % 24-hour survival % 28-day survival % Discharged neurologically intact % Shockable rhythms % Respiratory arrests % Nonshockable rhythms %

Before 31.5 52 61 48 48 3 9.5 20.5

After 18.4 63 58 40 42 3 3 13.4

P† ns p < .01 ns ns ns ns ns .04

* SD, standard deviation; ROSC, return of spontaneous circulation; ns, nonsignificant. † Fisher’s exact test.

period, in which residents round on patients, allowing a second opportunity to identify and call attention to patients who may require additional evaluation or resources. Reactions to team meetings have been generally positive, as suggested by survey responses (Appendix 1, available in online article). Adverse comments regarding the meeting have been infrequent but have included complaints that some team members arrive late and that the location should be changed.

Team Member Medical resident Medical interns (2) Step-down unit nurse ICU fellow Nurse practitioner Pharmacist Anesthesiology resident Respiratory therapist ICU attending Police

Role on Team Runs ACLS algorithms Chest compressions Injects medications, runs defibrillator Procedures, assists running code IV, IO access, other procedures Drug preparation, recordkeeping Airway, ventilation, procedures Airway, ventilation Maintain teamwork, debrief Crowd control, safety

* ICU attending can be called in but is generally absent at night. If the ICU fellow is out of house, the on-call ICU resident carries a pager. ACLS, advanced cardiovascular life support; IV, intravenous; IO, intraosseous.

SIMULATION TRAINING In developing our two-layer code response, we recognized the need to provide specific training, with an emphasis on their roles and our associated expectations. Previously, the hospital regarded ACLS or BLS training as both necessary and sufficient for the care of an arresting patient. We now regard these certification activities as key steps in knowledge acquisition but not themselves sufficient for optimal individual and team performance. In analyzing how the needs of the arresting patient could best be met, we developed three main training activities: (1) site-specific BLS training, (2) on-site ACLS refresher training (both now described), and (3) defibrillator training (described on page 162). 1. Site-Specific BLS Training. Ward nurses and related personnel are run through simulations in which they arrive at a patient’s room with medications and discover that the patient is not breathing and has no pulse. We typically find that, in the course of an hour of simulations, a team of nurses and nursing assistants progress, from being hesitant and slow in the delivery of CPR to being fast and crisp and aware of other personnel and the tasks that remain to be completed. Specific tasks for first and second responders are described and practiced. The training enables personnel to rehearse wheeling a crash cart into a room, calling for help, and providing key information to colleagues. We are currently working toward the goal of having all ward nurses complete this training module twice per year. 2. On-Site ACLS Refresher Training. In this course, which is repeated each month for rotating house staff serving as code team members, the most common ACLS events are reviewed in simulation scenarios. Because the medicine ward team on call forms the basic physician core of the arrest team, each of the five internal medicine teams spends an hour together for these exercises; the course takes the place of the medicine department’s

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

161

The Joint Commission Journal on Quality and Patient Safety “morning report” on the scheduled day. Ward nurses, code nurses, pharmacists, and respiratory therapists distribute themselves among the hour-long sessions. This course allows the team to work together in a less stressful environment and practice crisis resource management and the major ACLS protocols. Provider roles and the use of cognitive aids are emphasized throughout this course. In addition, we have identified principles of arrest management that can be applied to any arrest while the team members are arriving: ■ Check pulse first, observing breathing while you do this. ■ If pulseless, start CPR. Teammates should place backboard and prepare to take over CPR after two minutes. ■ While performing CPR, attach defibrillator pads (if not done already) and determine whether the patient has either a shockable or nonshockable arrest. ■ If RN enters room with CPR in progress, immediately prepare to inject epinephrine. ■ If pharmacist enters room with CPR in progress, don’t wait for leader to request preparation of epinephrine, just prepare epinephrine and announce it is ready. Our overall philosophy for all these courses is to treat them as analogous to sports training, in which the instructors’ role is to provide advice and coaching on how to perform better the next time. Repetitive practice is provided so that individual improvement can be demonstrated. Our language emphasizes a constant search for excellence rather than certification of competence. Existing work with pediatric resuscitation training supports the efficacy of this approach.36 The course has received uniformly high approval of all participants, as indicated by postcourse evaluations presented in Appendix 2 (available in online article).

IMPROVEMENT OF FAMILIARITY WITH EQUIPMENT: MASTERING THE DEFIBRILLATOR One of the events that led us to plan the CPR redesign initiative was a root cause analysis (prompted by a fatality) that revealed a widespread lack of knowledge among nurses and physicians regarding the defibrillator/pacer unit. Competency fairs and periodic manual testing of equipment failed to produce the type of hands-on expertise needed to recognize and work through technical and mechanical challenges posed by real arrests. In response, we set up a series of meetings with the manufacturer to answer the question, “What would a real expert know and do when using the defibrillator?” On the basis of these conversations and human factors issues, we made a number of key policy changes, including the following: ■ Establishment of a “ready mode” for defibrillator use, in which pulse oximetery sensors, electrocardiogram electrodes, and 162

defibrillation pads are attached and ready for use. ■ Purchase of new crash carts, on which the defibrillator would sit about 18 inches (45.7 centimeters) higher and have greater visibility throughout the room ■ Creation of a training program for nurses on the use of the defibrillator. Originally targeted for code team nurses, the program has expanded to emergency department and critical care nurses and anesthesiologists. Groups of four to five persons rotate through 30-minute sessions in which a mannequin simulator experiences a number of arrest scenarios designed to highlight a different area of knowledge or problem solving. Participants are expected to organize themselves as a team and provide CPR and ventilation while a teammate in the “hot seat” is responsible for managing the defibrillator. The specific scenarios and goals are described in Table 4 (page 163). Participants completing the training were highly receptive of the experience and its value, as indicated in Appendix 3 (available in online article). Responses were assessed for difference according to whether the participant had either greater than or less than 10 years’ clinical experience. Only one question revealed a difference according to experience, with less experienced participants having more difficulty (< 10 years, median = 8, versus 5 for > 10 years; p = .002 by MannWhitney U test for nonnormally distributed data).

IMPROVING KNOWLEDGE OF EQUIPMENT In the past, crash carts were opened up only during annual competency fairs, and the training pool of equipment consisted of outdated carts and equipment. Defibrillators used in hospital ACLS courses were not the ones used in the clinical environment. Given these impediments to high performance, we established the policy that real training required real equipment and that the equipment needed to be accessible at all times. A fully stocked crash cart was placed in a centralized skills laboratory open to code team members. The cart is moved from this location to other training activities as needed, but with the understanding among nurses, pharmacists, and house staff that it had a “home” that can be visited at any hour. Storage of drugs in the “pharmacy drawer” of the crash cart became problematic from an accreditation standpoint, so the standard pharmacy tray of the demonstration cart was replaced with a full-scale photograph of the drugs available. To further reinforce use of actual equipment in training during our monthly mock codes, we insist that participants use all equipment and services that they would in a normal patient encounter—including crash carts, defibrillators, and IV tubing. We use the open crash carts from these events as part of our “random audit” quality monitoring process to ensure that carts are properly stocked.

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

The Joint Commission Journal on Quality and Patient Safety Table 4. Patient and Equipment Scenarios Used in Defibrillator Training Course* Scenario 1 2 3 4

5

Situation and Setup Unmonitored VF arrest with no equipment problems Fine VF arrest; instructor accidentally hits “synchronize” button. Monitored VT/VF arrest; pads look like they are connected but they are not. VT/VF arrest lead II is displayed at start of scenario. Instructor dislodges electrode while performing CPR. Team takes over code after prolonged CPR. Presenting rhythm is asystole.

Learning Objectives 1. Identify rhythm and deliver counter shock. 2. Operate defibrillator without interruption in CPR. Understand that in fine VF, the defibrillator cannot find an R-wave and will not discharge a shock. 1. Use of paddles as an alternate means of checking rhythm and delivering therapy 2. Can also switch leads if 3-lead system is being used as well to monitor rhythms. 1. Understand that a “broken line” indicates a misconnection not asystole 2. Switch to lead I paddles/combo pads to continue management while lead is replaced. 1. Confirm all leads attached. 2. Demonstrate ability to switch leads. 3. Print strip in two different leads after calling off further resuscitation.

* Instructors playing the role of first responder orchestrate some of the technical problems shown. Responders in groups of two arrive and assume all duties, including cardiopulmonary resuscitation (CPR) and managing the defibrillator. VF, ventricular fibrillation; VT, ventricular tachycardia.

Subsequent to the change in philosophy, purchases of new crash carts and related equipment have included extra units for demonstration use or use in the simulation center. To further reinforce familiarity of crash cart contents, posters, such as the following, highlighted specific components and were placed in specific work areas: ■ Airway equipment and its location within the carts for respiratory department and anesthesia department call rooms ■ Posters with instructions of defibrillator/pacer placed in all medicine department team rooms and in intensive care areas ■ Posters with organization of drawers prepared for pharmacy department and ICU nurse conference rooms Additional posters were posted as follows: ■ Posters describing team organization placed at location of team meeting and in the ICU ■ Poster replicas of cognitive aids placed at team meeting locations and the simulation center ■ Posters on doors to indicate that there is a crash cart behind the door and that it is everyone’s job to be aware of where lifesaving equipment is located (Appendix 4, available in online article).

USE OF MOCK CODES TO PROVIDE SYSTEM OVERSIGHT Unannounced mock arrests are conducted on both regular wards and other locations on a monthly basis. Mock arrests involve a portable simulator, which is introduced to ward-based and other personnel with the directive that if “his condition seems to change,” go ahead and check pulses and breathing and treat the arrest as if it is real. Dispatch operators, code team members, and others (all initially unaware that the arrest is a simulation) respond as if it were a real event. In a mock arrest,

one of the authors, positioned at the door, states, “This is a drill, treat it as if were real.” Use of IV access, transcutaneous pacing, cardioversion, IV drugs, and so on, are given as if it were a human patient.35 In these activities, the emphasis is placed on the assessment of the entire cardiac arrest ecosystem rather than training per se, although both topics are addressed in a short debriefing that follows. In essence, the program reflects our desire to locate problems within the system prospectively and to make corrections proactively. We find this method of system analysis and continuous improvement more effective and efficient than the use of root cause analysis or Failure Mode and Effects Analysis. Although all techniques are complementary, the lifelike nature of in situ codes is probably only second to real-life near misses as a naturalistic method of error trapping. Nearly all modifications in crash cart contents, equipment on the wards, content of training activities, and policies and procedures have derived from observations made during mock arrests, and since their institution, there have not been any near misses or critical events concerning emergency cardiovascular care.

COGNITIVE AID PREPARATION AND DISTRIBUTION In all phases of education and workplace preparation, we emphasize the use of printed algorithms, dosing guides, and other checklists to ensure compliance with ACLS protocols. To facilitate use of the cognitive aids (for shockable and nonschockable pathways [Appendix 5, available in online article]), at the start of each month, cards are given out to all code respondents and to the house staff; extras are available at all training sessions. We suggest to team leaders that they use one trainees or other knowledgeable bystanders at a code to read the algorithm and remind the leader of the next steps. The flowchart diagram shown

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

163

The Joint Commission Journal on Quality and Patient Safety in Figure 1 is used in teaching to indicate the key decision point represented by “shockability” and the use of the cognitive aid for the applicable scenario.

EMPHASIS ON PREVENTION If a patent’s condition is allowed to deteriorate to the point of a cardiopulmonary arrest, the likelihood of meaningful survival is quite low. All CPR–related training emphasizes the greater benefit of intervening before an arrest. In our twice-daily team meetings, as stated, identification of patients who may be unstable or a source of concern represents a key item. Often, a visit to these patients and assessment of priorities is made following the team meeting, thereby, hopefully, preventing a cardiac arrest’s occurrence. Formal and informal analysis of arrests with ward staff and team members includes evaluation of code conduct, determination of the possible presence of warning signs before the arrest, and reiteration of the role of the hospital’s medical emergency team.

ADDITIONAL ELEMENTS 1. Event Review and Analysis. Ideally, all critical events are debriefed afterward, but often an arrest victim is transferred to the ICU for further resuscitation and stabilization, making this impractical. Therefore, in January 2009, we began weekly review of arrests and other critical events brought to our attention. Evaluation focuses on antecedent events and preventability of the arrest, management of the event itself, and system factors that may have contributed or detracted from success. Phone operators keep a log of place, time, and identity of each arresting patient. Information regarding arrest type, survival, and preexisting vital sign abnormalities is added to the demographic information. Rates of cardiac arrests and medical emergency team calls are computed on a ward-to-ward basis, with census data on bed occupancy as a denominator. Data are analyzed for long-term trends and other patterns and presented to nursing and medical leaders of the various wards. Long-term mortality is established later by computerized comparison of arrest and mortality data. 2. CPR Website. To keep the departments and members participating in arrest management up to date on new products and equipment, training course schedules, and advances in resuscitation, in January 2010 we created an internal website, which also features training videos; however, no information on specific patients, events, or trends is provided. 3. Increased On-Site ACLS Training. In the process of examining training practices, we found that many of the hospital’s personnel received ACLS training at centers outside our hospi164

tal, while our in-hospital courses were taught by outside instructors, often with equipment not used in our hospital. Accordingly, in mid-2011 ACLS instructors on the CPR Committee began an initiative to increase the number of in-house nurses trained as ACLS instructors, to increase the number of in-house courses, and to ensure that the same crash carts and defibrillators are used throughout the hospital. We have increased the number of in-house ACLS instructors from 4 to 20, with courses offered at least every month rather than a only a few times per year.

Discussion The various components of our hospital’s approach to CPR were examined and compared with an ideal model of resuscitation. Opportunities for improvement ranged from acquisition of better equipment to development of enhanced training programs that reinforce basic skills and teamwork and that provide a pathway toward mastery of advanced skills. A great deal of effort has been placed at investigating specific interventions that improve arrest outcomes,13 yet relatively little attention has been devoted to improving the system of care for resuscitation.37–41 In reducing our intervention to its essential components, we suggest that a mature hospital-based resuscitation system include the following components: ■ Data analysis and the ability to broadly define arrest trends and resuscitation needs ■ Development of local methods for approaching the arresting patient ■ Emphasis on prevention42 ■ Establishment of training programs reflecting the items above ■ System examination and oversight ■ Administrative processes that maximize interaction between all components Much of this list overlaps with an effort to describe successful rapid response systems.15 We recognize that no such consensus exists here but nonetheless propose this list as a means of stimulating conversation. Of all the components, we consider the last—administrative processes that maximize interaction between all components—to be the main characteristic of our system. We have created a general script for arrest management that is revisited and reinforced through a variety of activities and in multiple venues. All the programs described reinforce one another and involve integration at the level of personnel and message. Limitations of this system redesign mainly concern the effort required to maintain it while continuing its redesign. For example, we have been successful in creating a stable instructor base

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

The Joint Commission Journal on Quality and Patient Safety for fewer than 50% of our courses and programs, with the responsibility for teaching subsequently still largely resting on the authors. We cannot yet define how all the elements necessary for the system might be self-sustaining. House-staff duties related to code team membership are in constant competition with other clinical and educational responsibilities, so that maintaining proper training and readiness in teaching programs is a constant challenge. With reductions in work hours and greater reliance on shift work, getting this transient population trained and updated at the start of a rotation is increasingly difficult. Accordingly, more specialized functions such as defibrillator use and certain procedures (for example, echocardiogram and intraosseous needle and invasive line placement) have been shifted away from residents to smaller and more spatially stable groups such as ICU nurses and fellows. We also acknowledge the fact that the entirety of our redesign effort is based on a qualitative assessment of system function and need for improvement. We have not been able to advocate the use of the individual components on the basis of demonstratively reducing morbidity and mortality. We believe that our system better facilitates application of therapies that do save lives and provides an architecture for rapid implementation of new proven therapies. Data in Table 2 (page 161) and Appendix 6 (available in online article) indicate some improvement over baseline, but also demonstrate that we have not sustained an upward trend in survival after the first three years. These trends remind us that it is a constant challenge to keep a hospital prepared for generally rare events. Following institution of our rapid response system11 and the eTeam, our arrest numbers have been reduced significantly, and our survival frequencies are better than those reported from national databases, so these trends may reflect small-sample-size variation around an acceptable average. Hands-on CPR time, time to first shock, time to delivery of epinephrine, time to run through causes of pulseless electrical activity, and adherence to algorithms, among other metrics, have been studied by others. In the future we will monitor the effectiveness of the program by using standard reporting metrics, including return of spontaneous circulation, 24-hour survival, survival to discharge, and neurologic status. However, with the small number of arrest occurring in our institution, we will also monitor team performance (in particular, role clarity and adherence to ACLS guidelines) through monthly in situ simulated cardiac arrests-mock codes. There are numerous opportunities for improvement. As is typical for many university-based residencies, many house staff rotate among three different medical centers—each with different code team structures, equipment, and cultures and thereby

a lack of consensus on how to manage arrests. Hospital systems that do not house educational programs have fewer challenges with turnover and time constraints and are in a position to make significant advances in interprofessional education and team building not achievable with an academic staffing model.

Conclusion We have described the process and elements of an effort to improve the delivery of emergency cardiovascular care at our hospital. The interventions were developed on the basis of a large-scale view of the process of delivering BLS and ACLS, identification of key decision nodes and actions, and compartmentalization of specific functions in space and time and according to personnel. With training tailored to their respective roles, the individual providers have developed greater understanding of their function on the code team. The resuscitation system functions with an implicit requirement of self-examination and is constantly evolving. In our view, maximum performance is achieved through the dedication of individuals to the patient, and a system that promotes creativity about how to allow the code team’s members the greatest degree of success. J Geoffrey K. Lighthall, MD, PhD, is Associate Professor, Department of Anesthesia, Stanford University School of Medicine, Stanford, California; and the Department of Anesthesia, US Department of Veterans Affairs (VA) Palo Alto Health Care System, Palo Alto, California. Michael Mayette, MD, FRCPC, is Critical Care Medicine Fellow, Critical Care Medicine Program, Department of Medicine, Stanford University School of Medicine. T. Kyle Harrison, MD, is Clinical Assistant Professor (Affiliated), Department of Anesthesia, Stanford University School of Medicine, and VA Palo Alto Health Care System. Please address correspondence to Geoffrey K. Lighthall, [email protected].

Online-Only Content

8

See the online version of this article for

Appendix 1. Registered Nurses’ (N = 23) and Respiratory Therapists’ (N = 11) Responses to Survey on Attitudes Toward Daily Code Team Meetings, May 2011–June 2012 Appendix 2. Responses (N = 72) to Survey on the Refresher Advanced Cardiac Life Support Course, 2001–2009 Appendix 3. Nurses’ (N = 27) and Pharmacists’ (N = 4) Responses to Survey After First Iteration of Defibrillator Skills Course, 2007 Appendix 4. Door Poster Appendix 5. Two-Sided Cognitive Aid: Shockable and Nonshockable Pathways Appendix 6. Outcome Characteristics by Year, 2005–2011

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

165

The Joint Commission Journal on Quality and Patient Safety References 1. Ornato JP, et al. Impact of resuscitation system errors on survival from in-hospital cardiac arrest. Resuscitation. 2012;83(1):63–69. 2. Eisenberg MS, Mengert TJ. Cardiac resuscitation. N Engl J Med. 2001 Apr 26;344(17):1304–1313. 3. Peberdy MA, et al. Cardiopulmonary resuscitation of adults in the hospital: A report of 14720 cardiac arrests from the National Registry of Cardiopulmonary Resuscitation. Resuscitation. 2003;58(3):297–308. 4. Peberdy MA, et al. Survival from in-hospital cardiac arrest during nights and weekends. JAMA. 2008 Feb 20;299(7):785–792. 5. Nadkarni VM, et al. First documented rhythm and clinical outcome from inhospital cardiac arrest among children and adults. JAMA. 2006 Jan 4;295(1):50–57. 6. DeVita MA, et al. “Identifying the hospitalised patient in crisis”—A consensus conference on the afferent limb of rapid response systems. Resuscitation. 2010;81(4):375–382. 7. Abella BS, et al. Chest compression rates during cardiopulmonary resuscitation are suboptimal: A prospective study during in-hospital cardiac arrest. Circulation. 2005 Feb 1;111(4):428–434. 8. Chan PS, et al. Delayed time to defibrillation after in-hospital cardiac arrest. N Engl J Med. 2008 Jan 3;358(1):9–17. 9. Hunziker S, et al. Human factors in resuscitation: Lessons learned from simulator studies. J Emerg Trauma Shock. 2010;3(4):389–394. 10. Aufderheide TP, Lurie KG. Death by hyperventilation: A common and lifethreatening problem during cardiopulmonary resuscitation. Crit Care Med. 2004;32(9 Suppl):S345–351. 11. Lighthall GK, et al. Introduction of a rapid response system at a United States Veterans Affairs hospital reduced cardiac arrests. Anesth Analg. 2010;111(3):679–686. 12. Morley PT, et al. Part 3: Evidence evaluation process: 2010 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science with Treatment Recommendations. Circulation. 2010 Oct 19;122(16 Suppl 2):S283-290. 13. Nadkarni VM, et al. Part 2: International collaboration in resuscitation science: 2010 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science with Treatment Recommendations. Circulation. 2010 Oct 19;122(16 Suppl 2):S276–82. 14. Morrison LJ, et al. Part 8: Advanced life support: 2010 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science with Treatment Recommendations. Circulation. 2010 Oct 19;122(16 Suppl 2):S345–421. 15. Devita MA, et al. Findings of the first consensus conference on medical emergency teams. Critical Care Med. 2006;34(9):2463–2478. 16. Iwami T, et al. Effectiveness of bystander-initiated cardiac-only resuscitation for patients with out-of-hospital cardiac arrest. Circulation. 2007 Dec 18;116(25):2900–2907. 17. Ong ME, et al. Comparison of chest compression only and standard cardiopulmonary resuscitation for out-of-hospital cardiac arrest in Singapore. Resuscitation. 2008;78(2):119–126. 18. Bhende MS, Karasic DG, Karasic RB. End-tidal carbon dioxide changes during cardiopulmonary resuscitation after experimental asphyxial cardiac arrest. Am J Emerg Med. 1996;14(4):349–350. 19. Garnett AR, et al. End-tidal carbon dioxide monitoring during cardiopulmonary resuscitation. JAMA. 1987 Jan 23–30;257(4):512–515. 20. Pokorná M, et al. A sudden increase in partial pressure end-tidal carbon dioxide (P(ET)CO(2)) at the moment of return of spontaneous circulation. J Emerg Med. 2010;38(5):614–621.

166

21. Heidenreich JW, et al. Single-rescuer cardiopulmonary resuscitation: “Two quick breaths”—An oxymoron. Resuscitation. 2004;62(3):283–289. 22. Heidenreich JW, et al. Uninterrupted chest compression CPR is easier to perform and remember than standard CPR. Resuscitation. 2004;63(2):123–130. 23. Menegazzi JJ, et al. Combination pharmacotherapy with delayed countershock vs standard advanced cardiac life support after prolonged ventricular fibrillation. Prehosp Emerg Care. 2000;4(1):31–37. 24. Niemann JT, et al. Immediate countershock versus cardiopulmonary resuscitation before countershock in a 5-minute swine model of ventricular fibrillation arrest. Ann Emerg Med. 2000;36(6):543–546. 25. Yakaitis RW, et al. Influence of time and therapy on ventricular defibrillation in dogs. Crit Care Med. 1980;8(3):157–163. 26. Marsch SC, et al. Performance of first responders in simulated cardiac arrests. Crit Care Med. 2005;33(5):963–967. 27. ECC Committee, American Heart Association. 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2005 Dec 13;112(24 Suppl):IV1–203. 28. Hazinski MF, et al. Major changes in the 2005 AHA Guidelines for CPR and ECC: Reaching the tipping point for change. Circulation. 2005 Dec 13;112(24 Suppl):IV206–211. 29. Sayre MR, et al. Part 5: Adult basic life support: 2010 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science with Treatment Recommendations. Circulation. 2010 Oct 19;122(16 Suppl 2):S298–324. 30. Neumar RW, et al. Implementation strategies for improving survival after out-of-hospital cardiac arrest in the United States: Consensus recommendations from the 2009 American Heart Association Cardiac Arrest Survival Summit. Circulation. 2011 Jun 21;123(24):2898–2910. 31. Hayes CW, et al. Residents feel unprepared and unsupervised as leaders of cardiac arrest teams in teaching hospitals: A survey of internal medicine residents. Crit Care Med. 2007;35(7):1668–1672. 32. Hunziker S, et al. Brief leadership instructions improve cardiopulmonary resuscitation in a high-fidelity simulation: A randomized controlled trial. Crit Care Med. 2010;38(4):1086–1091. 33. Hunziker S, et al. Teamwork and leadership in cardiopulmonary resuscitation. J Am Coll Cardiol. 2011 Jun 14;57(24):2381–2388. 34. Hunziker S, et al. Hands-on time during cardiopulmonary resuscitation is affected by the process of teambuilding: A prospective randomised simulatorbased trial. BMC Emerg Med. 2009 Feb 14;9:3. 35. Lighthall GK, Poon T, Harrison TK. Using in situ simulation to improve in-hospital cardiopulmonary resuscitation. Jt Comm J Qual Patient Saf. 2010;36(5):209–216. 36. Sutton RM, et al. Low-dose, high-frequency CPR training improves skill retention of in-hospital pediatric providers. Pediatrics. 2011;128(1):e145–151. 37. Chan PS, Nallamothu BK. Improving outcomes following in-hospital cardiac arrest: Life after death. JAMA. 2012 May 9;307(18):1917–1918. 38. Davis DP. The ART of resuscitation. A new program for cardiopulmonary arrest calls. JEMS. 2010;35(9):48–49. 39. Seethala RR, Esposito EC, Abella BS. Approaches to improving cardiac arrest resuscitation performance. Curr Opin Crit Care. 2010;16(3):196–202. 40. Soar J, Edelson DP, Perkins GD. Delivering high-quality cardiopulmonary resuscitation in-hospital. Curr Opin Crit Care. 2011;17(3):225–230. 41. Sutton RM, Nadkarni V, Abella BS. “Putting it all together” to improve resuscitation quality. Emerg Med Clin North Am. 2012;30(1):105–122. 42. Smith GB. In-hospital cardiac arrest: Is it time for an in-hospital “chain of prevention”? Resuscitation. 2010;81(9):1209–11.

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

The Joint Commission Journal on Quality and Patient Safety Online-Only Content

8

Appendix 1. Registered Nurses’ (N = 23) and Respiratory Therapists’ (N = 11) Responses to Survey on Attitudes Toward Daily Code Team Meetings, May 2011–June 2012* Statement The meetings help me recognize my role and contribution to the team. The meetings help me recognize the roles and contributions of others. The meetings likely lead to better team performance. The meetings help provide ongoing education in resuscitation. We should keep the team meetings.

Median Response, (1–5) Scale† RNs RTs 4 4 4 5 4.5 4 5 4 5 5

* Median responses of nonnormally distributed data are shown. None of the differences between the RNs and RTs (respiratory therapists) were significant (Fisher’s exact test). †

A score of 1 is the most negative response (“Disagree strongly”) and 5 is the most positive response (“Agree strongly”) on the 5-point Likert scale.

Appendix 2. Responses (N = 72) to Survey on the Refresher Advanced Cardiac Life Support Course, 2001–2009* Survey Question Rated by 5-Point Likert Scale of 1 (“Disagree strongly”) to 5 (“Agree strongly”) 1. I am more aware of my role on the cardiac arrest team. 2. I am more aware of the role of others on the cardiac arrest team. 3. I am more comfortable performing my role on the cardiac arrest team. 4. I feel more comfortable being the leader of a cardiac arrest team. (All respondents) Residents only (Designated code leaders) Nonresident participants 5. The instructional content of this course was appropriate for my level of training. Rated by 5-point Likert Scale of 1 (“Others better”) to 5 (“This program is best”) 6. Compared to code team preparation at other institutions where you practice, how do you rate the efforts at the VA Palo Alto?

Median 4 4 5 4 4† 3 5

5

* Median responses of nonnormally distributed data are shown. VA, US Department of Veterans Affairs. Versus median for nonresidents; p < .001; Mann-Whitney U test. “Resident” was defined as postgraduate year 2–3 internal medicine residents), and “nonresidents” as interns, nurses, and pharmacists. †

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

AP1

The Joint Commission Journal on Quality and Patient Safety Online-Only Content

8

Appendix 3. Nurses’ (N = 27) and Pharmacists’ (N = 4) Responses to Survey After First Iteration of Defibrillator Skills Course, 2007* Survey Question 1 (“Others better”) to 10 (“This course better”): Relative to other courses or modules on this subject, how does this course compare? How does simulation training compare overall to other training modalities you have experienced? As far as knowledge acquisition goes, how does this course compare to other experiences? As far as potential for knowledge retention goes, how does this compare to other experiences?

Median

15%–75% CI

8 9 8 9

8–9 8–10 8–9 8–9

1 (“No impact”) to 10 (“Best ever”): Did this course increase your overall confidence in your role as a resuscitation team member? Do you think this course will help you function better as a code team member?

9 9

7–9 8–10

1 (“Too easy), 5 (“Just right’), 10 (“Too difficult”): How does the material covered compare with your current knowledge base?†

6

5–8

1 (”Would rather have others”) to 10 (“Outstanding”): How do you perceive the credibility and overall knowledge of your instructors?

10

9–10

1 (“Disagree”) to 10 (“Absolutely agree”): I believe this course will lead to improvements in the quality and safety of patient care.

10

9–10

*A 1–10 Likert scale was used, wherein 1 indicates the lowest level of agreement and 10 the greatest. The 31 respondents were in practice a mean of 11.84 years (standard error, 1.49), and 55% of them had advanced cardiac life support certification. CI, confidence interval. † Participants with less experience (< 10 years) showed a median of 8 versus 5 (data not shown) for those with more experience (> 10 years), p = .002 by MannWhitney U test for nonnormally distributed data.

AP2

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

The Joint Commission Journal on Quality and Patient Safety Online-Only Content

Appendix 4. Door Poster

These posters, placed on outside doors of rooms containing a crash cart, are intended to raise everyone’s awareness as to the whereabouts of emergency equipment.

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

AP3

The Joint Commission Journal on Quality and Patient Safety Online-Only Content

8

Appendix 5. Two-Sided Cognitive Aid: Shockable and Nonshockable Pathways

This laminated cognitive aid is distributed widely throughout the hospital. One side indicates the treatment of a shockable rhythm, and the opposite side shows the nonshockable pathway.

AP4

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

The Joint Commission Journal on Quality and Patient Safety Online-Only Content

8 Appendix 6. Outcome Characteristics by Year, 2005–2011

Top graph: The percentage of arrests with return of spontaneous circulation (ROSC), 24-hour survival, 28-day survival, and discharge (DC) to home or nursing home with intact neurologic function is shown. The data for 2005 and 2006 predate the majority of the interventions described in this article. Bottom graph: Arrests are broken down into categories of respiratory arrests (Resp)-, shockable-, and nonshockable-presenting rhythms, to indicate differences in survival. All patients with ROSC were admitted to the ICU.

April 2013 Volume 39 Number 4 Copyright 2013 © The Joint Commission

AP5