- Email: [email protected]
Root Cause Analysis of ICU Adverse Events in the Veterans Health Administration
ARTICLE IN PRESS The Joint Commission Journal on Quality and Patient Safety 2017; ■■:■■–■■
Root Cause Analysis of ICU Adverse Events in the Veterans Health Administration Gregory S. Corwin, MPH; Peter D. Mills, PhD, MS; Hasan Shanawani, MD; Robin R. Hemphill, MD; MPH
Background: ICUs’ provision of complex care for critically ill patients results in an environment with a high potential for adverse events. A study was conducted to characterize adverse events in Veterans Health Administration (VHA) ICUs that underwent root cause analysis (RCA) and to identify the root causes and their recommended actions. Methods: This retrospective observational study of RCA reports concerned events that occurred in VHA ICUs or as a result of ICU processes from January 1, 2013, through December 31, 2014. The type of event, root causes, and recommended actions were measured. Results: Some 70 eligible RCAs were identified in 47 of the 120 facilities with an ICU in the VHA system. Delays in care (30.0%) and medication errors (28.6%) were the most common types of events. There were 152 root causes and 277 recommended actions. Root causes often involved rules, policies, and procedure processes (28.3%), equipment/supply issues (15.8%), and knowledge deficits/education (15.1%). Common actions recommended were policy, procedure, and process actions (34.4%) and training/education actions (31.4%). Of the actions implemented, 84.4% had a reported effectiveness of “much better” or “better.” Conclusion: ICU adverse events often had several root causes, with protocols and process-of-care issues as root causes regardless of event type. Actions often included standardization of processes and training/education. Several recommendations can be made that may improve patient safety in the ICU, such as standardization of care process, implementation of team training programs, and simulation-based training.
I
n 1999 the Institute of Medicine (IOM) highlighted the problem of adverse events in health care, suggesting that up to 100,000 deaths in the United States were the result of preventable adverse events.1 Estimates based on more recent data now suggest that more than 400,000 deaths may be associated with preventable harm.2 During the last 15 years, reducing adverse events, and, more broadly, the development of strong patient safety systems, has become an increasingly important focus for health care organizations. The ICU is a complex environment, providing care for critically ill and unstable patients undergoing multiple medical interventions. Consequently, the ICU is an environment in which there may be an increased risk for adverse events. Much of the literature on ICU adverse events concerns their frequency in the ICU and the breakdown in individual types of events.3–8 The first article to report on ICU adverse events, appeared in 1980; Abramson et al. identified 145 adverse events in 4,720 admissions in a five-year period.3 Most recent studies have found a higher prevalence of ICU adverse events, with reports of up to 100 adverse events per 1,000 patient-days.4–8 It has been reported that 20%–30% of patients admitted to the ICU experience an adverse event,4,5,7 with severity of illness, intensity of care, and time of ICU exposure as all predictive of ICU events.7 Similarly, increased ICU workload and utilization have been suggested to increase medical errors.9 1553-7250/$-see front matter Published by Elsevier Inc. on behalf of The Joint Commission. http://dx.doi.org/10.1016/j.jcjq.2017.04.009
Rothschild et al., in a prospective observational study, found that 45% of adverse events were preventable and that 13% of these preventable events were life-threatening or fatal.5 Similarly, in a study of Spanish ICUs, 58% of patients had an incident, of which roughly a third could be classified as an adverse event.10 Others have reported that having two or more adverse events during an ICU admission was an independent risk factor for mortality.4 Not only do errors affect patients in the ICU, but preventable iatrogenic events occur in up to 27% of patients before ICU admission, 80% of which contribute to the need for ICU admission.11 Yet although studies report a high number of adverse events, they do not report on the specific cause for these events or on actions taken to mitigate future risk. Root cause analysis (RCA), a process to review and identify contributing factors to adverse events or close calls,12,13 been used in the Veterans Health Administration (VHA) system since 2000.13 In addition, the US Department of Veterans Affairs (VA) National Center for Patient Safety (NCPS) established an RCA database,13 which most facilities were using by 2000, to provide a resource to review and research RCA data and which has been used to analyze adverse events in a range of clinical settings.13–17 The objective of this study was to characterize reported ICU adverse events occurring in the VHA system using RCA reports. In particular, we identified the root causes and the actions taken to address these events. Our intent was to identify whether there were general factors across the different categories of adverse events.
ARTICLE IN PRESS 2
Gregory S. Corwin, MPH, et al
METHODS VA National Center for Patient Safety RCA Program
The VA NCPS was established in 1998 as a dedicated center responsible for leading patient safety efforts throughout the VHA system in response to the VA’s Patient Safety Improvement Initiative.13 The VHA system of 144 medical centers is the largest integrated health care system in the United States, providing comprehensive care to more than 8.9 million Veterans every year.18 NCPS takes a systems-level approach to patient safety, focusing on the underlying causes of adverse events and corrective actions to address the causes, placing an emphasis on learning rather than blaming individuals.13 Currently, VHA facilities perform RCA reviews of adverse events and close calls. 12,13 RCAs focus on determining what happened, why it happened, and what can be done to prevent it from happening again.13 The emphasis of RCA is on systems-level issues that can be addressed to prevent future adverse events in a nonpunitive manner.13 RCA case data provide a unique perspective for analyzing adverse events because, in addition to identifying the type of adverse event that occurred, the data identify the causes for adverse events and actions taken in response to these events. VHA RCA Process
When an adverse event or close call is reported in a VHA facility, the patient safety manager (PSM) scores the event using the Safety Assessment Code (SAC) Matrix (Sidebar 1).12 The SAC matrix allows PSMs to assess an event on both its severity and its probability of occurrence.13 Using this matrix, the PSM determines an actual and potential SAC score for a given case. Scores range from 1 through 3, with a score of 1 representing the lowest severity and probability of reoccurrence and a score of 3 representing the highest severity and probability of reoccurrence. Events with a SAC score of 3 require an RCA, which must be completed within 45
RCA of ICU Adverse Events
days. After the RCA is complete, the information from the case is entered into the NCPS RCA database. An RCA can either be performed as an individual RCA for a single event or as an aggregated review that analyzes groups of similar incidents to identify common causes for a given event type. Study Design and Ethics
Our study was observational, with a retrospective look at all RCA reports of adverse events occurring in ICUs across the VHA system from January 1, 2013, through December 31, 2014. The study had Institutional Review Board approval and approval from the Research and Development Committee at the White River Junction VA Medical Center (Committee for the Protection of Human Subjects at Dartmouth College #2169). RCA Case Search and Inclusion Criteria
We identified all RCA cases in which an adverse event occurred in the ICU for 2013 and 2014. Potentially eligible RCA cases were identified by searching for cases that used ICU location codes and cases that had the ICU listed as an event location. To be eligible for inclusion, an adverse event had to have occurred either in the ICU or as a result of ICU processes. An adverse event was defined (based on the NCPS definition) as “untoward incidents, therapeutic misadventures, iatrogenic injuries, or other adverse occurrences directly associated with care or services provided within the jurisdiction of a medical facility, outpatient clinic, or other VHA facility.”12(p. 2) In addition, events in which errors reached the patient, but no harm was reported, were included as an adverse event. Halted RCA cases or aggregated reviews were excluded. Variables
A codebook was developed to review and code the RCA cases. The main variables were type of event, root causes, and
Sidebar 1. Safety Assessment Code (SAC) Matrix Severity & Probability
Catastrophic
Major
Moderate
Minor
Frequent—Likely to occur immediately or within a short period (may happen several times in 1 year) Occasional—Probably will occur (may happen several times in 1 to 2 years) Uncommon—Possible to occur (may happen sometime in 2 to 5 years) Remote—Unlikely to occur (may happen sometime in 5 to 30 years)
3
3
2
1
3 3 3
2 2 2
1 1 1
1 1 1
Matrix score: 3 = highest risk; 2 = intermediate risk; 1 = lowest risk. Catastrophic—Death or permanent loss of function not related to the natural course of the patient’s illness or underlying condition. Major—Permanent lessening of bodily functioning not related to the natural course of the patient’s illness or underlying condition. Moderate—Increased length of stay or increased level of care for one or two patients. Minor—No injury, nor increased length of stay nor increased level of care. Adapted from US Department of Veterans Affairs, Veterans Health Administration. VHA National Patient Safety Improvement Handbook. Mar 4, 2011. Accessed Jun 18, 2017. http://www.va.gov/vhapublications/ViewPublication.asp?pub_ID=2389.
ARTICLE IN PRESS Volume ■■, No. ■■, ■■ 2017
3
Table 1. ICU Adverse Events for Calendar Year (CY) 2013 and CY 2014 by Type of Event and Definitions for Types of Adverse Events (N = 70) Type of Event
Percentage of Reported Events (%)*
Delay in care Medication
30.0 28.6
Medical procedure Equipment failure
11.4 8.6
Other Removal of lines, catheters, tubes, drains Transfusion
5.7 4.3
Elopement Discharge Suicide attempt
1.4 1.4 1.4
Fall
1.4
Airway/ventilation
1.4
Infection
0
4.3
Definition An event in which a patient does not receive timely care for his or her illness. An event in which there were errors in the preparation and/or the administration of medication to a patient. This includes wrong dose, wrong route, wrong medication, contraindications, allergies, wrong patient, etc. An event in which there were errors during the performance of a procedure. An event in which medical equipment does not function appropriately due to technical issues. Any other events not in the above categories. An event in which a patient’s line, catheter, tube, or drain becomes accidentally dislodged or disconnected. An event in which there were errors in the preparation and/or the administration of blood products to a patient. An event in which a patient leaves the medical facility without authorization to leave. An event resulting from the improper discharge of a patient. An event in which a patient attempts a nonfatal, self-directed, potentially injurious behavior with an intent to die as a result of the behavior; might not result in injury. An event in which a patient comes to rest inadvertently on the ground or floor or lower level. An event in which the airway and/or ventilation of a patient was not properly maintained. This includes inadvertent extubations, inappropriate extubations, displacement of endotracheal tubes, etc. Any infection acquired during the course of care at a medical center.
*Because of rounding, adds up to only 99.9%.
actions. Patient characteristics such as age and gender were collected if provided. The type of event was classified into 13 categories (Table 1). Root causes and actions were coded based on those identified by the local RCA team. The root cause categories were knowledge deficit/education; rules, policies, procedure processes; general communication; equipment and supply issues; lack of patient information; staffing; cultural; other; and no root causes. The action categories were training and education actions; policy, procedure, and processes actions; communication actions; equipment actions; documentation actions; staffing actions; and other. Both root causes and actions were coded further into more specific subcategories (Table 2 and Table 3). Action effectiveness was also coded on the basis of the reported effectiveness (in terms of an actions outcome measure) of a given action by the local site where the adverse event occurred. The effectiveness was coded as much better; better; same; worse; not reported; not implemented; not measured; and no action to measure. If an action had the effectiveness reported at multiple time points, the most recent time point was coded and used for the analysis. Data Processing
Using the codebook developed, 10 RCA cases were coded independently by two authors [G.S.C., P.D.M.], and a Cohen’s kappa coefficient was calculated to measure the
interrater agreement between the two coders (kappa = 0.966). The remaining cases were coded independently. RESULTS RCA Case Eligibility
Our search of the RCA database identified 115 potentially eligible cases (Figure 1). We excluded 14 cases because the Flow Diagram of Included RCA Cases of Reported ICU Adverse Events
Figure 1: The search of the root cause analysis (RCA) database identified 115 potentially eligible cases, as shown; 70 cases met all eligibility criteria and were included in the analysis. CY, calendar year.
ARTICLE IN PRESS 4
Gregory S. Corwin, MPH, et al
RCA of ICU Adverse Events
Table 2. General Categories and Subcategories for Root Causes General Root Causes Knowledge Deficit/Education Rules, Policies, Procedure Processes
General Communication Equipment and Supply Issues
Lack of Patient Information
Staffing Cultural
Other No Root Cause
Subcategory Root Causes Knowledge deficit related to something not used or performed often Lack of policies Lack of standardized processes for procedures Policies or procedure processes need improvement Policies or procedure processes not followed and/or performed correctly Communication between departments Communication within a department/team Equipment not functioning and/or used properly Equipment and/or supplies missing/unavailable Human factors engineering Environmental issues Lack of documentation Lack of standardization of documentation Lack of real-time or updated patient information Lack of standardization in the ordering process Error in use of ordering systems No subcategories Culture did not promote speaking up Overreliance on others and/or equipment Culture did not promote teamwork No subcategories No subcategories
Table 3. General Categories and Subcategories for Actions General Actions Training and Education Actions
Policy, Procedure, and Process Actions
Communication Actions Equipment Actions
Documentation Actions Staffing Actions Other
Subcategory Actions Development of cognitive aids, signs in clinical environment, checklists Update of required competencies Online training modules and/or training videos Simulation-based training Standardization of clinical processes and procedures Review and/or develop protocols/processes Create new policy Update/improve policy No subcategories New equipment, upgrade/replace equipment, or remove old equipment Standardize equipment and supplies Redesign of work area/environment Develop or improve ordering forms No subcategories No subcategories
RCA was either halted or incomplete. Of the remaining 101 cases, 31 were excluded because the event described did not fit our definition of an adverse event or the adverse event did not occur in the ICU. A total of 70 cases met all eligibility criteria and were included in the analysis. Facility and Patient Characteristics
The 70 cases occurred in 47 of the 120 facilities with an ICU in the VHA system. Of the 70 cases, 60 (85.7%) were in highcomplexity facilities, with the remaining cases occurring in medium- or low-complexity facilities. Generally, high-complexity facilities, here consisting of three complexity groups (highest
complexity, high complexity, and mid-high complexity) are characterized by medium to high patient volume, medium- to high-risk patients, and large research and teaching programs; medium-complexity facilities are characterized by medium patient volume, low-risk patients, and small research and teaching programs; and low-complexity facilities are characterized by low patient volume, low-risk patients, and no research and teaching programs. During the study period, there were 870,436 ICU bed-days of care across all 120 VHA facilities with an ICU. The average patient age was 66.1 years (standard deviation, 14.2). Of the 43 RCA cases that reported gender, 39 (90.7%) of the patients were male (Table 4).
ARTICLE IN PRESS Volume ■■, No. ■■, ■■ 2017
5
Table 4. Characteristics of Patients Experiencing a Reported ICU Adverse Event (N = 70) Characteristic Average age* (SD) Gender† (%) Male Female
CY 2014 (n = 31)
CY 2013 (n = 39)
Total (N = 70)
67.8 (13.0)
66.4 (15.5)
66.1 (14.2)
89.5 10.5
91.6 8.3
90.7 9.3
*Not reported for 32 patients. † Not reported for 27 patients. CY, calendar year; SD, standard deviation.
Types of Adverse Events
The most common types of events were delays in care (30.0%) and medication errors (28.6%) (Table 1). Examples of delay in care events were delays in transfer to the ICU typically centered on bed availability, delays in providing appropriate treatment, and delays in diagnosing a patient’s changing condition. Most medication errors involved the administration of the wrong dose of a medication or infusing a medication at an incorrect rate. The next most common events were medical procedure errors (11.4%) and equipment failures (8.6%). Medical procedure errors generally involved misplacing lines, while equipment failures often involved the failure of alarm/alert mechanisms. Root Causes
There were 152 root causes identified (Figure 2). Root causes involving rules, policies, and procedure processes (28.3%) were
the most common. Other root causes identified were equipment and supply issues (15.8%), knowledge deficit/education (15.1%), general communication (12.5%), and lack of patient information (11.8%). Root causes associated with hospital culture, staffing, and “other” made up the remaining root causes identified. One case did not identify a root cause. Root Causes by Type of Adverse Event
Root causes were evaluated for the most common types of events (Table 5). Rules, policies, and procedure processes issues were consistently identified root causes across the most common events. Events identifying communication issues and equipment/supply availability issues as a root cause were often delay in care events. Medication errors typically reported general knowledge deficits or a knowledge deficit related to not using or performing a procedure or treatment often enough. Equipment issues pertaining to human factors engineering—which addresses how people interact with equipment and systems 19 —were common root causes for medication errors, procedure errors, and equipment failures. Actions
There were 277 recommended actions identified (Figure 3). Policy, procedure, and process actions (34.4%) and training and education actions (31.4%) were the most common. Other recommended actions were equipment actions (13.0%) and documentation actions (12.3%). The remaining actions were communication actions, staffing actions, and other actions. One case did not recommend any actions.
Root Causes for Reported ICU Adverse Events for CY 2013 and CY 2014 (N = 152)
Figure 2: Of the 152 root causes identified, those involving rules, policies, and procedure processes (28.3%) were the most common. CY, calendar year.
ARTICLE IN PRESS 6
Gregory S. Corwin, MPH, et al
RCA of ICU Adverse Events
Table 5. Most Common Root Causes and Recommended Actions for the Main Types of Reported ICU Adverse Events for Calendar Year (CY) 2013 and CY 2014 Type of Event
Root Causes
Actions
Delay in care (Root cause N = 45; Action N = 76)
1. Rules, policies, and procedure processes (26.7%; n = 12) – Lack of standardized processes for procedures 2. General communication (22.2%; n = 10) – Communication between departments and within teams 3. Other (15.6%; n = 7) 4. Equipment and supply issues (13.3%; n = 6) – Equipment and/or supplies missing/unavailable 1. Rules, policies, and procedure processes (25.0%; n = 10) – Lack of policies – Lack of standardized processes for procedures 2. Knowledge deficit/education (22.5%; n = 9) – General knowledge deficit (lack of education/lack of training) – Deficit related to something not used or performed often 3. Equipment and supply issues (20.0%; n = 8) – Human factors engineering 4. Lack of patient information (15.0%; n = 6) – Updated patient information, documentation issues, ordering issues 1. Rules, policies, and procedure processes (35.3%; n = 6) – Policies or procedure processes not followed properly – Lack of standardized processes for procedures 2. Equipment and supply issues (17.6%; n = 3) – Human factors engineering 3. Knowledge deficit/education (17.6%; n = 3) – General knowledge deficit (lack of education/lack of training) 1. Rules, policies, and procedure processes (26.7%; n = 4) – Lack of standardized processes for procedures 2. Equipment and supply issues (20.0%; n = 3) – Human factors engineering
1. Policy, procedure, and process actions (35.5%; n = 27) – Standardization of clinical processes and procedures – Review and/or develop protocols/processes 2. Training and education actions (30.3%; n = 23) – General development of training and educational programs and materials – Development of cognitive aids, signs, checklists – Simulation-based training 3. Equipment actions (11.8%; n = 9) – New equipment, upgrade equipment, or remove equipment 1. Policy, procedure, and process actions (36.4%; n = 24) – Standardization of clinical processes and procedures – Review and/or develop protocols/processes 2. Training and education actions (31.8%; n = 21) – General development of training and educational programs and materials – Development of cognitive aids, signs, checklists 3. Documentation actions (19.7%; n = 13) – Develop or improve ordering forms
Medication (Root cause N = 40; Action N = 66)
Procedure (Root cause N = 17; Action N = 22)
Equipment failure (Root cause N = 15; Action N = 43)
1. Training and education actions (45.5%; n = 10) – General development of training and educational programs and materials – Simulation-based training 2. Policy, procedure, and process actions (27.3%; n = 6) – Standardization of clinical processes and procedures 3. Equipment actions (18.2%; n = 4) – New equipment, upgrade equipment, or remove equipment 1. Training and education actions (27.9%; n = 12) – General development of training and educational programs and materials – Development of cognitive aids, signs, checklists 2. Equipment actions (25.6%; n = 11) – New equipment, upgrade equipment, or remove equipment 3. Policy, procedure, and process actions (23.3%; n = 10) – Standardization of clinical processes and procedures – Review and/or develop protocols/processes – Create new policy
Actions by Type of Adverse Event
Action Effectiveness
Actions were evaluated for the most common types of events (Table 5). Consistent across events was the recommendation for standardization of clinical processes and procedures. Simulation-based training actions were recommended for delay in care events and procedure errors. Actions involving developing cognitive aids, signs, and checklists were recommended for delay in care events, medication errors, and equipment failures.
Of the 277 recommended actions, 270 (97.5%) were implemented and had a reported effectiveness (Figure 4). Of the 270 implemented actions, 47.4% had a reported effectiveness of “much better,” 37.0% had a reported effectiveness of “better,” 7.4% had a reported effectiveness of “same,” and the remaining 8.2% were either “not reported” or “not measured.” Reported effectiveness was also evaluated for each action category (Table 6). Within each action category, a majority of the
ARTICLE IN PRESS Volume ■■, No. ■■, ■■ 2017
7
Recommended Actions for Reported ICU Adverse Events for CY 2013 and CY 2014 (N = 277)
Figure 3: Of the 277 recommended actions identified, policy, procedure, and process actions (34.4%) and training and education actions (31.4%) were the most common. CY, calendar year.
Reported Effectiveness of Implemented Actions (N = 270)
Figure 4: Of the 277 recommended actions, 270 (97.5%) were implemented and had a reported effectiveness.
actions implemented had a reported effectiveness of “much better” or “better.” DISCUSSION
The goal of our study was to identify common underlying factors that are involved across the different categories of reported adverse events. The study is unique in that it is the first to systematically explore reported ICU adverse events using a national RCA database, enabling us to examine the
root causes identified for reported adverse events, as well as recommendations for improvements. We were able to identify root causes that were repeatedly identified across multiple categories of adverse events. The types of adverse events we identified in the RCA database are consistent with the types of adverse events that have been identified in previous studies.4,6–8 A majority of the events had either an actual or potential SAC score of 3. This is not surprising, given that events with a SAC score
ARTICLE IN PRESS 8
Gregory S. Corwin, MPH, et al
RCA of ICU Adverse Events
Table 6. Reported Action Effectiveness by Action Categories Reported Effectiveness (%) Action Category Training and Education Actions (N = 87) Policy, Procedure, and Process Actions (N = 94)† Communication Actions (N = 13) Equipment Actions (N = 34)† Documentation Actions (N = 31) Staffing Actions (N = 9) Other (N = 2)
Much Better/ Better Same Worse Other* 85.06
4.60
0
10.34
81.91
9.57
0
8.51
84.62
7.69
0
7.69
82.35
11.76
0
5.88
96.77
3.23
0
0
66.67 100
11.11 0
0 0
22.22 0
*Not Reported; Not Measured. † Because of rounding, adds up to only 99.99%.
of 3 require an RCA and also represent events that have a higher severity and probability of reoccurrence. Most of the reported adverse events identified more than one root cause. This observation is consistent with Reason’s model of organizational safety suggesting that, rather than a single factor, multiple factors have to align for an adverse event to occur.20 However, the relative contribution of any individual factor is difficult to ascertain. Across most of the reported adverse events, a lack of standardized processes and/or protocols was identified as a root cause. Similarly, standardization of clinical processes and procedures was a common recommended action. These observations are consistent with the recent emphasis on the importance of standardized critical care protocols.21 For example, care bundles specifically directed toward a disease process of care have been successful in improving clinical outcomes in the critically ill.22 Similarly, in a prospective study, the implementation of a resident handoff bundle resulted in a significant decrease in medical errors and preventable adverse events in hospitalized children.23 It has been demonstrated that just having protocols does not necessarily result in the protocols being followed or improved outcomes.24 Additional factors such as protocol implementation, compliance, and behavior change may play a role in the success of ICU protocols.24–26 One study of ICU patients reported that just over 50% of eligible patients were receiving care consistent with best-practice treatment guidelines.27 Similarly, in a trauma population, almost 50% of the patients had care with moderate to severe deviation in compliance with best-practice guidelines;28 greater compliance was associated with improved clinical outcomes. Getting physicians to follow protocols and/or guidelines has been shown to be challenging, with multiple barriers
interfering with physician behavior change and adherence with guidelines.29 Effective teamwork, which has been found to be important in overcoming barriers to adhering to protocols and processes of care in the ICU,30,31 involves team communication, team coordination, team decision making, and team leadership.31 A number of interventions have been used to improve team performance.30 For example, implementing a mandatory checklist of protocols and care objectives on rounds was effective in adherence to best practice.32 In an attempt to improve communication and teamwork, the VHA Medical Team Training (MTT) program was developed.33 MTT was based on crew resource management (CRM), with the goal of increasing the understanding of safety threats in the acute care setting.33 The MTT program has been successfully implemented in the VHA system and has been demonstrated to improve patient outcomes and improve teamwork climate.33–35 Similarly, a meta-analysis of 20 studies that examined CRM in acute hospital settings demonstrated a positive response from participants and an increase in their knowledge and behaviors; however, there was insufficient evidence to determine the impact on patient outcomes.36 Almost all the reported adverse events had actions that were intended to address at least one root cause through a training and education action. A variety of training and education actions were recommended, such as simulationbased training; online training modules; updating required competencies; and the addition of cognitive aids, signs, and checklists in the clinical environment. Although training actions are typically identified as weak actions, the literature suggests that simulation-based training may be the best approach.37–42 Simulation-based training has been shown to be effective in improving technical and nontechnical skills and in turn patient safety.43 A study by Paull et al. demonstrated that simulation-based team training in the perioperative setting improves teamwork and communication among staff.41 It should be noted that not all actions are equal. When viewed individually, some actions are considered to be stronger, and other actions weaker. NCPS developed an action hierarchy, with actions categorized in terms of their relative strength.42 Stronger actions are recognized as those that standardize or simplify processes, while weaker actions are recognized as those involving general training or new procedures and policies.42 In looking at adverse drug events, Mills et al. found that actions that focused on bedside clinical care improvements were associated with improved outcomes.44 However, although general training and education have been identified as weaker actions, more immersive forms of training and education such as MTT and simulation-based training have been shown to be effective in improving clinical care.33–35,37,39,41 Individual sites implemented and reported the effectiveness for more than 97% of the recommended actions.
ARTICLE IN PRESS Volume ■■, No. ■■, ■■ 2017
Most of the reported adverse events were addressed by several actions. Overall, 84.4% of the actions that were implemented resulted in a reported effectiveness of either “much better” or “better.” Despite the self-reported nature of this information, the findings do suggest that the actions recommended and subsequently implemented by sites were effective. In addition, consistent across the different action categories, the majority of actions for a given action category had a reported effectiveness of “much better” or “better.” On the basis of our findings, we suggest the following recommendations for improvement of patient safety in the ICU: • Development and implementation of standardized protocols and processes of care, including strategies for education and compliance with protocols • Development and implementation of team training programs designed using CRM concepts • Simulation-based training for improvement of technical and nontechnical skills Limitations
This study had several limitations. The unique patient population and organizational structure of the VHA may limit the generalizability of the findings to medical centers outside the VHA. The primary limitation, however, is the potential for the underreporting of adverse events. One of the reasons there may be underreporting is that the RCA system relies on reporting by frontline personnel. Different providers may have different views on whether a case is in fact an adverse event. For example, one study comparing adverse event reporting rates of ICU staff to that of observers found that the observers’ reporting rate was nearly double that of the ICU staff.45 In addition, because of the complexity of ICU care, adverse events may be missed when they are instead viewed as a complication of care. We also explored whether facility complexity levels between the facilities in which the RCA cases included in our study occurred and the facilities that did not have RCA cases, to find that there were no significant differences. In addition to underreporting, some adverse events that are reported may not be reviewed using an individual RCA. Individual RCAs are mandatory only for events with SAC scores of 3. For events with a SAC score < 3, individual hospitals decide whether or not to perform an RCA. Although hospital resources and priorities could affect this decision, SAC scores are intended to help hospitals identify and direct their resources to serious events even if they are near misses. An adverse event may also be addressed in an aggregate review of several adverse events that occur throughout a given year at a single site. Moreover, RCA is only one tool available for the review of adverse events; events with lower SAC scores may be addressed by using other patient safety tools. Given the reported frequency of ICU adverse events, particularly those that are life-threatening or fatal, there was a relatively
9
low number of reported ICU adverse events that resulted in an RCA. Although it is unclear whether the number of RCAs is the “correct” number, the fact that the types of events captured in the RCAs corresponds to the types of events reported in the literature suggests that the included RCAs captured the significant adverse events occurring in the ICU.4,6,8 Our analysis of a given action’s effectiveness is also limited. Action effectiveness was self-reported by each site, so we could not evaluate how a given site determined the effectiveness of a given action. In addition, although each action’s effectiveness was reported, the assessment of a single action’s effectiveness might have reflected its implementation in combination with several other actions. This may explain why actions that are noted as being weaker are reported to have a “much better” or “better” effectiveness after being implemented. Since health care’s adoption of the RCA process in the 1990s, RCA itself has come under increased scrutiny. Recently, questions have been raised about the general effectiveness of the RCA process in health care, and recommendations have been suggested to improve it.46–49 Although the VA RCA process provides valuable information on adverse events, there are areas for improvement, particularly regarding follow-up (implemented actions and the sharing of lessons learned). Most important, RCA is most effective when used as part of a set of tools, such as failure mode and effects analysis (FMEA), which have received much less attention in health care—and are used in the VA system, although to a lesser degree.47,50 Efforts are under way to make the FMEA process easier and more accessible to VA medical centers, as well as to make the RCA process more widely available with results that are more “SMART” (specific, measurable, achievable, relevant, time-sensitive). Finally, the importance of sharing lessons learned across VA medical centers, which is a key priority, cannot be overemphasized. CONCLUSION
Reported ICU adverse events typically had several root causes, with protocol and process-of-care issues (lack of standardized approaches) consistently identified as problems. Actions often included standardization of processes and the development and implementation of appropriate training and education. Acknowledgments. This material is the result of work supported with resources and the use of facilities at the US Department of Veterans Affairs (VA) National Center for Patient Safety Field Office in White River Junction, Vermont, and the VA National Center for Patient Safety in Ann Arbor, Michigan. Disclaimer. The views expressed in this article do not necessarily represent the views of the US Department of Veterans Affairs or of the United States Government. Presentations. An abstract of this work was presented at the American Thoracic Society (ATS) International Conference, May 19–24, 2017, Washington, DC. Conflicts of Interest. All authors report no conflicts of interest.
ARTICLE IN PRESS 10
Gregory S. Corwin, MPH, et al
Gregory S. Corwin, MPH, formerly Patient Safety Fellow, US Department of Veterans Affairs (VA) National Center for Patient Safety Field Office, White River Junction, Vermont, is Medical Student, University of Arkansas for Medical Sciences College of Medicine, Little Rock. Peter D. Mills, PhD, MS, is Director, VA National Center for Patient Safety Field Office, White River Junction, Vermont, and Adjunct Associate Professor, Department of Psychiatry, Geisel School of Medicine at Dartmouth, Hanover, New Hampshire. Hasan Shanawani, MD, is Patient Safety Physician, VA National Center for Patient Safety, Ann Arbor, Michigan, and Clinical Instructor, Division of Pulmonary and Critical Care Medicine, University of Michigan, Ann Arbor. Robin R. Hemphill, MD, MPH, is Veterans Health Administration Chief Patient Safety and Risk Awareness Officer, and Director, VA National Center for Patient Safety, Ann Arbor. Please address correspondence to Peter D. Mills, [email protected].
REFERENCES 1. Institute of Medicine. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Press, 2000. 2. James JT. A new, evidence-based estimate of patient harms associated with hospital care. J Patient Saf. 2013;9:122–128. 3. Abramson NS, et al. Adverse occurrences in intensive care units. JAMA. 1980 Oct 3;244:1582–1584. 4. Garrouste-Orgeas M, et al. Selected medical errors in the intensive care unit: results of the IATROREF study: parts I and II. Am J Respir Crit Care Med. 2010 Jan 15;181:134–142. 5. Rothschild JM, et al. The Critical Care Safety Study: the incidence and nature of adverse events and serious medical errors in intensive care. Crit Care Med. 2005;33:1694–1700. 6. Thomas AN, Taylor RJ. Review of patient safety incidents reported from critical care units in North-West England in 2009 and 2010. Anaesthesia. 2012;67:706–713. 7. Valentin A, et al. Patient safety in intensive care: results from the multinational Sentinel Events Evaluation (SEE) study. Intensive Care Med. 2006;32:1591–1598. 8. Welters ID, et al. Major sources of critical incidents in intensive care. Crit Care. 2011;15:R232. 9. Steyrer J, et al. Attitude is everything? The impact of workload, safety climate, and safety tools on medical errors: a study of intensive care units. Health Care Manage Rev. 2013;38:306–316. 10. Merino P, et al. Adverse events in Spanish intensive care units: the SYREC study. Int J Qual Health Care. 2012;24:105– 113. 11. Garry DA, et al. A prospective multicentre observational study of adverse iatrogenic events and substandard care preceding intensive care unit admission (PREVENT). Anaesthesia. 2014;69:137–142. 12. US Department of Veterans Affairs, Veterans Health Administration. VHA National Patient Safety Improvement Handbook. Mar 4, 2011. Accessed Jun 18, 2017. http:// www.va.gov/vhapublications/ ViewPublication.asp?pub_ID=2389. 13. Bagian JP, et al. Developing and deploying a patient safety program in a large health care delivery system: you can’t fix what you don’t know about. Jt Comm J Qual Improv. 2001;27:522–532. 14. Giardina TD, et al. Root cause analysis reports help identify common factors in delayed diagnosis and treatment of outpatients. Health Aff (Millwood). 2013;32:1368–1375. 15. Lee A, et al. Root cause analysis of serious adverse events among older patients in the Veterans Health Administration. Jt Comm J Qual Patient Saf. 2014;40:253–262. 16. Miller KE, et al. Wrong-side thoracentesis: lessons learned from root cause analysis. JAMA Surg. 2014;149:774–779.
RCA of ICU Adverse Events 17. Mills PD, et al. Suicide attempts and completions in Veterans Affairs nursing home care units and long-term care facilities: a review of root-cause analysis reports. Int J Geriatr Psychiatry. 2016;31:518–525. 18. US Department of Veterans Affairs. Veterans Health Administration: About VHA. (Updated: Mar 9, 2017.) Accessed Jun 18, 2017. http://www.va.gov/health/aboutVHA .asp. 19. Salvendy G, editor. Handbook of Human Factors and Ergonomics, 3rd ed. Hoboken, NJ: John Wiley & Sons, 2006. 20. Reason J. Human error: models and management. BMJ. 2000 Mar 18;320:768–770. 21. Weled BJ, et al. Critical care delivery: the importance of process of care and ICU structure to improved outcomes: an update from the American College of Critical Care Medicine Task Force on Models of Critical Care. Crit Care Med. 2015;43:1520–1525. 22. Dellinger RP, et al. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med. 2013;39:165–228. 23. Starmer AJ, et al. Rates of medical errors and preventable adverse events among hospitalized children following implementation of a resident handoff bundle. JAMA. 2013 Dec 4;310:2262–2270. 24. Sevransky JE, et al. Protocols and hospital mortality in critically ill patients: the United States Critical Illness and Injury Trials Group Critical Illness Outcomes Study. Crit Care Med. 2015;43:2076–2084. 25. Bonello RS, et al. An intensive care unit quality improvement collaborative in nine Department of Veterans Affairs hospitals: reducing ventilator-associated pneumonia and catheterrelated bloodstream infection rates. Jt Comm J Qual Patient Saf. 2008;34:639–645. 26. Centofanti JE, et al. Use of a daily goals checklist for morning ICU rounds: a mixed-methods study. Crit Care Med. 2014; 42:1797–1803. 27. Ilan R, et al. Knowledge translation in critical care: factors associated with prescription of commonly recommended best practices for critically ill patients. Crit Care Med. 2007;35:1696–1702. 28. Rice TW, et al. Deviations from evidence-based clinical management guidelines increase mortality in critically injured trauma patients. Crit Care Med. 2012;40:778–786. 29. Cabana MD, et al. Why don’t physicians follow clinical practice guidelines? A framework for improvement. JAMA. 1999 Oct 20;282:1458–1465. 30. Dietz AS, et al. A systematic review of teamwork in the intensive care unit: what do we know about teamwork, team tasks, and improvement strategies? J Crit Care. 2014;29:908– 914. 31. Reader TW, et al. Developing a team performance framework for the intensive care unit. Crit Care Med. 2009; 37:1787–1793. 32. Byrnes MC, et al. Implementation of a mandatory checklist of protocols and objectives improves compliance with a wide range of evidence-based intensive care unit practices. Crit Care Med. 2009;37:2775–2781. 33. Dunn EJ, et al. Medical team training: applying crew resource management in the Veterans Health Administration. Jt Comm J Qual Patient Saf. 2007;33:317–325. 34. Neily J, et al. Association between implementation of a medical team training program and surgical mortality. JAMA. 2010 Oct 20;304:1693–1700. 35. Carney BT, et al. Improving perceptions of teamwork climate with the Veterans Health Administration medical
ARTICLE IN PRESS Volume ■■, No. ■■, ■■ 2017
36.
37.
38.
39.
40.
41.
team training program. Am J Med Qual. 2011;26:480– 484. O’Dea A, O’Connor P, Keogh I. A meta-analysis of the effectiveness of crew resource management training in acute care domains. Postgrad Med J. 2014;90:699– 708. Barsuk JH, et al. Simulation-based mastery learning reduces complications during central venous catheter insertion in a medical intensive care unit. Crit Care Med. 2009;37:2697– 2701. Gold JA, et al. Integrating the electronic health record into high-fidelity interprofessional intensive care unit simulations. J Interprof Care. 2015;29:562–563. Gundrosen S, Solligård E, Aadahl P. Team competence among nurses in an intensive care unit: the feasibility of in situ simulation and assessing non-technical skills. Intensive Crit Care Nurs. 2014;30:312–317. Hebbar KB, et al. Simulation-based paediatric intensive care unit central venous line maintenance bundle training. Intensive Crit Care Nurs. 2015;31:44–50. Paull DE, et al. The effect of simulation-based crew resource management training on measurable teamwork and communication among interprofessional teams caring for postoperative patients. J Contin Educ Nurs. 2013;44:516–524.
11 42. US Department of Veterans Affairs (VA) National Center for Patient Safety. Root Cause Analysis Tools. Washington, DC: VA, 2015. 43. Naik VN, Brien SE. Review article: simulation: a means to address and improve patient safety. Can J Anaesth. 2013;60: 192–200. 44. Mills PD, et al. Effective interventions and implementation strategies to reduce adverse drug events in the Veterans Affairs (VA) system. Qual Saf Health Care. 2008;17:37–46. 45. Capuzzo M, et al. Reporting of unintended events in an intensive care unit: comparison between staff and observer. BMC Emerg Med. 2005 May 27;5:3. 46. Card AJ. The problem with ‘5 whys.’ BMJ Qual Saf. Epub 2016 Sep 2. 47. Peerally MF, et al. The problem with root cause analysis. BMJ Qual Saf. 2017;26:417–422. 48. Wu AW, Lipshutz AK, Pronovost PJ. Effectiveness and efficiency of root cause analysis in medicine. JAMA. 2008 Feb 13;299:685–687. 49. National Patient Safety Foundation (NPSF). RCA2: Improving Root Cause Analyses and Actions to Prevent Harm. Boston: NPSF, 2015. 50. Dixon-Woods M, et al. Safer Clinical Systems: Evaluation Findings. London: The Health Foundation, 2014.
Root Cause Analysis of ICU Adverse Events in the Veterans Health Administration Gregory S. Corwin, MPH; Peter D. Mills, PhD, MS; Hasan Shanawani, MD; Robin R. Hemphill, MD; MPH
Background: ICUs’ provision of complex care for critically ill patients results in an environment with a high potential for adverse events. A study was conducted to characterize adverse events in Veterans Health Administration (VHA) ICUs that underwent root cause analysis (RCA) and to identify the root causes and their recommended actions. Methods: This retrospective observational study of RCA reports concerned events that occurred in VHA ICUs or as a result of ICU processes from January 1, 2013, through December 31, 2014. The type of event, root causes, and recommended actions were measured. Results: Some 70 eligible RCAs were identified in 47 of the 120 facilities with an ICU in the VHA system. Delays in care (30.0%) and medication errors (28.6%) were the most common types of events. There were 152 root causes and 277 recommended actions. Root causes often involved rules, policies, and procedure processes (28.3%), equipment/supply issues (15.8%), and knowledge deficits/education (15.1%). Common actions recommended were policy, procedure, and process actions (34.4%) and training/education actions (31.4%). Of the actions implemented, 84.4% had a reported effectiveness of “much better” or “better.” Conclusion: ICU adverse events often had several root causes, with protocols and process-of-care issues as root causes regardless of event type. Actions often included standardization of processes and training/education. Several recommendations can be made that may improve patient safety in the ICU, such as standardization of care process, implementation of team training programs, and simulation-based training.
I
n 1999 the Institute of Medicine (IOM) highlighted the problem of adverse events in health care, suggesting that up to 100,000 deaths in the United States were the result of preventable adverse events.1 Estimates based on more recent data now suggest that more than 400,000 deaths may be associated with preventable harm.2 During the last 15 years, reducing adverse events, and, more broadly, the development of strong patient safety systems, has become an increasingly important focus for health care organizations. The ICU is a complex environment, providing care for critically ill and unstable patients undergoing multiple medical interventions. Consequently, the ICU is an environment in which there may be an increased risk for adverse events. Much of the literature on ICU adverse events concerns their frequency in the ICU and the breakdown in individual types of events.3–8 The first article to report on ICU adverse events, appeared in 1980; Abramson et al. identified 145 adverse events in 4,720 admissions in a five-year period.3 Most recent studies have found a higher prevalence of ICU adverse events, with reports of up to 100 adverse events per 1,000 patient-days.4–8 It has been reported that 20%–30% of patients admitted to the ICU experience an adverse event,4,5,7 with severity of illness, intensity of care, and time of ICU exposure as all predictive of ICU events.7 Similarly, increased ICU workload and utilization have been suggested to increase medical errors.9 1553-7250/$-see front matter Published by Elsevier Inc. on behalf of The Joint Commission. http://dx.doi.org/10.1016/j.jcjq.2017.04.009
Rothschild et al., in a prospective observational study, found that 45% of adverse events were preventable and that 13% of these preventable events were life-threatening or fatal.5 Similarly, in a study of Spanish ICUs, 58% of patients had an incident, of which roughly a third could be classified as an adverse event.10 Others have reported that having two or more adverse events during an ICU admission was an independent risk factor for mortality.4 Not only do errors affect patients in the ICU, but preventable iatrogenic events occur in up to 27% of patients before ICU admission, 80% of which contribute to the need for ICU admission.11 Yet although studies report a high number of adverse events, they do not report on the specific cause for these events or on actions taken to mitigate future risk. Root cause analysis (RCA), a process to review and identify contributing factors to adverse events or close calls,12,13 been used in the Veterans Health Administration (VHA) system since 2000.13 In addition, the US Department of Veterans Affairs (VA) National Center for Patient Safety (NCPS) established an RCA database,13 which most facilities were using by 2000, to provide a resource to review and research RCA data and which has been used to analyze adverse events in a range of clinical settings.13–17 The objective of this study was to characterize reported ICU adverse events occurring in the VHA system using RCA reports. In particular, we identified the root causes and the actions taken to address these events. Our intent was to identify whether there were general factors across the different categories of adverse events.
ARTICLE IN PRESS 2
Gregory S. Corwin, MPH, et al
METHODS VA National Center for Patient Safety RCA Program
The VA NCPS was established in 1998 as a dedicated center responsible for leading patient safety efforts throughout the VHA system in response to the VA’s Patient Safety Improvement Initiative.13 The VHA system of 144 medical centers is the largest integrated health care system in the United States, providing comprehensive care to more than 8.9 million Veterans every year.18 NCPS takes a systems-level approach to patient safety, focusing on the underlying causes of adverse events and corrective actions to address the causes, placing an emphasis on learning rather than blaming individuals.13 Currently, VHA facilities perform RCA reviews of adverse events and close calls. 12,13 RCAs focus on determining what happened, why it happened, and what can be done to prevent it from happening again.13 The emphasis of RCA is on systems-level issues that can be addressed to prevent future adverse events in a nonpunitive manner.13 RCA case data provide a unique perspective for analyzing adverse events because, in addition to identifying the type of adverse event that occurred, the data identify the causes for adverse events and actions taken in response to these events. VHA RCA Process
When an adverse event or close call is reported in a VHA facility, the patient safety manager (PSM) scores the event using the Safety Assessment Code (SAC) Matrix (Sidebar 1).12 The SAC matrix allows PSMs to assess an event on both its severity and its probability of occurrence.13 Using this matrix, the PSM determines an actual and potential SAC score for a given case. Scores range from 1 through 3, with a score of 1 representing the lowest severity and probability of reoccurrence and a score of 3 representing the highest severity and probability of reoccurrence. Events with a SAC score of 3 require an RCA, which must be completed within 45
RCA of ICU Adverse Events
days. After the RCA is complete, the information from the case is entered into the NCPS RCA database. An RCA can either be performed as an individual RCA for a single event or as an aggregated review that analyzes groups of similar incidents to identify common causes for a given event type. Study Design and Ethics
Our study was observational, with a retrospective look at all RCA reports of adverse events occurring in ICUs across the VHA system from January 1, 2013, through December 31, 2014. The study had Institutional Review Board approval and approval from the Research and Development Committee at the White River Junction VA Medical Center (Committee for the Protection of Human Subjects at Dartmouth College #2169). RCA Case Search and Inclusion Criteria
We identified all RCA cases in which an adverse event occurred in the ICU for 2013 and 2014. Potentially eligible RCA cases were identified by searching for cases that used ICU location codes and cases that had the ICU listed as an event location. To be eligible for inclusion, an adverse event had to have occurred either in the ICU or as a result of ICU processes. An adverse event was defined (based on the NCPS definition) as “untoward incidents, therapeutic misadventures, iatrogenic injuries, or other adverse occurrences directly associated with care or services provided within the jurisdiction of a medical facility, outpatient clinic, or other VHA facility.”12(p. 2) In addition, events in which errors reached the patient, but no harm was reported, were included as an adverse event. Halted RCA cases or aggregated reviews were excluded. Variables
A codebook was developed to review and code the RCA cases. The main variables were type of event, root causes, and
Sidebar 1. Safety Assessment Code (SAC) Matrix Severity & Probability
Catastrophic
Major
Moderate
Minor
Frequent—Likely to occur immediately or within a short period (may happen several times in 1 year) Occasional—Probably will occur (may happen several times in 1 to 2 years) Uncommon—Possible to occur (may happen sometime in 2 to 5 years) Remote—Unlikely to occur (may happen sometime in 5 to 30 years)
3
3
2
1
3 3 3
2 2 2
1 1 1
1 1 1
Matrix score: 3 = highest risk; 2 = intermediate risk; 1 = lowest risk. Catastrophic—Death or permanent loss of function not related to the natural course of the patient’s illness or underlying condition. Major—Permanent lessening of bodily functioning not related to the natural course of the patient’s illness or underlying condition. Moderate—Increased length of stay or increased level of care for one or two patients. Minor—No injury, nor increased length of stay nor increased level of care. Adapted from US Department of Veterans Affairs, Veterans Health Administration. VHA National Patient Safety Improvement Handbook. Mar 4, 2011. Accessed Jun 18, 2017. http://www.va.gov/vhapublications/ViewPublication.asp?pub_ID=2389.
ARTICLE IN PRESS Volume ■■, No. ■■, ■■ 2017
3
Table 1. ICU Adverse Events for Calendar Year (CY) 2013 and CY 2014 by Type of Event and Definitions for Types of Adverse Events (N = 70) Type of Event
Percentage of Reported Events (%)*
Delay in care Medication
30.0 28.6
Medical procedure Equipment failure
11.4 8.6
Other Removal of lines, catheters, tubes, drains Transfusion
5.7 4.3
Elopement Discharge Suicide attempt
1.4 1.4 1.4
Fall
1.4
Airway/ventilation
1.4
Infection
0
4.3
Definition An event in which a patient does not receive timely care for his or her illness. An event in which there were errors in the preparation and/or the administration of medication to a patient. This includes wrong dose, wrong route, wrong medication, contraindications, allergies, wrong patient, etc. An event in which there were errors during the performance of a procedure. An event in which medical equipment does not function appropriately due to technical issues. Any other events not in the above categories. An event in which a patient’s line, catheter, tube, or drain becomes accidentally dislodged or disconnected. An event in which there were errors in the preparation and/or the administration of blood products to a patient. An event in which a patient leaves the medical facility without authorization to leave. An event resulting from the improper discharge of a patient. An event in which a patient attempts a nonfatal, self-directed, potentially injurious behavior with an intent to die as a result of the behavior; might not result in injury. An event in which a patient comes to rest inadvertently on the ground or floor or lower level. An event in which the airway and/or ventilation of a patient was not properly maintained. This includes inadvertent extubations, inappropriate extubations, displacement of endotracheal tubes, etc. Any infection acquired during the course of care at a medical center.
*Because of rounding, adds up to only 99.9%.
actions. Patient characteristics such as age and gender were collected if provided. The type of event was classified into 13 categories (Table 1). Root causes and actions were coded based on those identified by the local RCA team. The root cause categories were knowledge deficit/education; rules, policies, procedure processes; general communication; equipment and supply issues; lack of patient information; staffing; cultural; other; and no root causes. The action categories were training and education actions; policy, procedure, and processes actions; communication actions; equipment actions; documentation actions; staffing actions; and other. Both root causes and actions were coded further into more specific subcategories (Table 2 and Table 3). Action effectiveness was also coded on the basis of the reported effectiveness (in terms of an actions outcome measure) of a given action by the local site where the adverse event occurred. The effectiveness was coded as much better; better; same; worse; not reported; not implemented; not measured; and no action to measure. If an action had the effectiveness reported at multiple time points, the most recent time point was coded and used for the analysis. Data Processing
Using the codebook developed, 10 RCA cases were coded independently by two authors [G.S.C., P.D.M.], and a Cohen’s kappa coefficient was calculated to measure the
interrater agreement between the two coders (kappa = 0.966). The remaining cases were coded independently. RESULTS RCA Case Eligibility
Our search of the RCA database identified 115 potentially eligible cases (Figure 1). We excluded 14 cases because the Flow Diagram of Included RCA Cases of Reported ICU Adverse Events
Figure 1: The search of the root cause analysis (RCA) database identified 115 potentially eligible cases, as shown; 70 cases met all eligibility criteria and were included in the analysis. CY, calendar year.
ARTICLE IN PRESS 4
Gregory S. Corwin, MPH, et al
RCA of ICU Adverse Events
Table 2. General Categories and Subcategories for Root Causes General Root Causes Knowledge Deficit/Education Rules, Policies, Procedure Processes
General Communication Equipment and Supply Issues
Lack of Patient Information
Staffing Cultural
Other No Root Cause
Subcategory Root Causes Knowledge deficit related to something not used or performed often Lack of policies Lack of standardized processes for procedures Policies or procedure processes need improvement Policies or procedure processes not followed and/or performed correctly Communication between departments Communication within a department/team Equipment not functioning and/or used properly Equipment and/or supplies missing/unavailable Human factors engineering Environmental issues Lack of documentation Lack of standardization of documentation Lack of real-time or updated patient information Lack of standardization in the ordering process Error in use of ordering systems No subcategories Culture did not promote speaking up Overreliance on others and/or equipment Culture did not promote teamwork No subcategories No subcategories
Table 3. General Categories and Subcategories for Actions General Actions Training and Education Actions
Policy, Procedure, and Process Actions
Communication Actions Equipment Actions
Documentation Actions Staffing Actions Other
Subcategory Actions Development of cognitive aids, signs in clinical environment, checklists Update of required competencies Online training modules and/or training videos Simulation-based training Standardization of clinical processes and procedures Review and/or develop protocols/processes Create new policy Update/improve policy No subcategories New equipment, upgrade/replace equipment, or remove old equipment Standardize equipment and supplies Redesign of work area/environment Develop or improve ordering forms No subcategories No subcategories
RCA was either halted or incomplete. Of the remaining 101 cases, 31 were excluded because the event described did not fit our definition of an adverse event or the adverse event did not occur in the ICU. A total of 70 cases met all eligibility criteria and were included in the analysis. Facility and Patient Characteristics
The 70 cases occurred in 47 of the 120 facilities with an ICU in the VHA system. Of the 70 cases, 60 (85.7%) were in highcomplexity facilities, with the remaining cases occurring in medium- or low-complexity facilities. Generally, high-complexity facilities, here consisting of three complexity groups (highest
complexity, high complexity, and mid-high complexity) are characterized by medium to high patient volume, medium- to high-risk patients, and large research and teaching programs; medium-complexity facilities are characterized by medium patient volume, low-risk patients, and small research and teaching programs; and low-complexity facilities are characterized by low patient volume, low-risk patients, and no research and teaching programs. During the study period, there were 870,436 ICU bed-days of care across all 120 VHA facilities with an ICU. The average patient age was 66.1 years (standard deviation, 14.2). Of the 43 RCA cases that reported gender, 39 (90.7%) of the patients were male (Table 4).
ARTICLE IN PRESS Volume ■■, No. ■■, ■■ 2017
5
Table 4. Characteristics of Patients Experiencing a Reported ICU Adverse Event (N = 70) Characteristic Average age* (SD) Gender† (%) Male Female
CY 2014 (n = 31)
CY 2013 (n = 39)
Total (N = 70)
67.8 (13.0)
66.4 (15.5)
66.1 (14.2)
89.5 10.5
91.6 8.3
90.7 9.3
*Not reported for 32 patients. † Not reported for 27 patients. CY, calendar year; SD, standard deviation.
Types of Adverse Events
The most common types of events were delays in care (30.0%) and medication errors (28.6%) (Table 1). Examples of delay in care events were delays in transfer to the ICU typically centered on bed availability, delays in providing appropriate treatment, and delays in diagnosing a patient’s changing condition. Most medication errors involved the administration of the wrong dose of a medication or infusing a medication at an incorrect rate. The next most common events were medical procedure errors (11.4%) and equipment failures (8.6%). Medical procedure errors generally involved misplacing lines, while equipment failures often involved the failure of alarm/alert mechanisms. Root Causes
There were 152 root causes identified (Figure 2). Root causes involving rules, policies, and procedure processes (28.3%) were
the most common. Other root causes identified were equipment and supply issues (15.8%), knowledge deficit/education (15.1%), general communication (12.5%), and lack of patient information (11.8%). Root causes associated with hospital culture, staffing, and “other” made up the remaining root causes identified. One case did not identify a root cause. Root Causes by Type of Adverse Event
Root causes were evaluated for the most common types of events (Table 5). Rules, policies, and procedure processes issues were consistently identified root causes across the most common events. Events identifying communication issues and equipment/supply availability issues as a root cause were often delay in care events. Medication errors typically reported general knowledge deficits or a knowledge deficit related to not using or performing a procedure or treatment often enough. Equipment issues pertaining to human factors engineering—which addresses how people interact with equipment and systems 19 —were common root causes for medication errors, procedure errors, and equipment failures. Actions
There were 277 recommended actions identified (Figure 3). Policy, procedure, and process actions (34.4%) and training and education actions (31.4%) were the most common. Other recommended actions were equipment actions (13.0%) and documentation actions (12.3%). The remaining actions were communication actions, staffing actions, and other actions. One case did not recommend any actions.
Root Causes for Reported ICU Adverse Events for CY 2013 and CY 2014 (N = 152)
Figure 2: Of the 152 root causes identified, those involving rules, policies, and procedure processes (28.3%) were the most common. CY, calendar year.
ARTICLE IN PRESS 6
Gregory S. Corwin, MPH, et al
RCA of ICU Adverse Events
Table 5. Most Common Root Causes and Recommended Actions for the Main Types of Reported ICU Adverse Events for Calendar Year (CY) 2013 and CY 2014 Type of Event
Root Causes
Actions
Delay in care (Root cause N = 45; Action N = 76)
1. Rules, policies, and procedure processes (26.7%; n = 12) – Lack of standardized processes for procedures 2. General communication (22.2%; n = 10) – Communication between departments and within teams 3. Other (15.6%; n = 7) 4. Equipment and supply issues (13.3%; n = 6) – Equipment and/or supplies missing/unavailable 1. Rules, policies, and procedure processes (25.0%; n = 10) – Lack of policies – Lack of standardized processes for procedures 2. Knowledge deficit/education (22.5%; n = 9) – General knowledge deficit (lack of education/lack of training) – Deficit related to something not used or performed often 3. Equipment and supply issues (20.0%; n = 8) – Human factors engineering 4. Lack of patient information (15.0%; n = 6) – Updated patient information, documentation issues, ordering issues 1. Rules, policies, and procedure processes (35.3%; n = 6) – Policies or procedure processes not followed properly – Lack of standardized processes for procedures 2. Equipment and supply issues (17.6%; n = 3) – Human factors engineering 3. Knowledge deficit/education (17.6%; n = 3) – General knowledge deficit (lack of education/lack of training) 1. Rules, policies, and procedure processes (26.7%; n = 4) – Lack of standardized processes for procedures 2. Equipment and supply issues (20.0%; n = 3) – Human factors engineering
1. Policy, procedure, and process actions (35.5%; n = 27) – Standardization of clinical processes and procedures – Review and/or develop protocols/processes 2. Training and education actions (30.3%; n = 23) – General development of training and educational programs and materials – Development of cognitive aids, signs, checklists – Simulation-based training 3. Equipment actions (11.8%; n = 9) – New equipment, upgrade equipment, or remove equipment 1. Policy, procedure, and process actions (36.4%; n = 24) – Standardization of clinical processes and procedures – Review and/or develop protocols/processes 2. Training and education actions (31.8%; n = 21) – General development of training and educational programs and materials – Development of cognitive aids, signs, checklists 3. Documentation actions (19.7%; n = 13) – Develop or improve ordering forms
Medication (Root cause N = 40; Action N = 66)
Procedure (Root cause N = 17; Action N = 22)
Equipment failure (Root cause N = 15; Action N = 43)
1. Training and education actions (45.5%; n = 10) – General development of training and educational programs and materials – Simulation-based training 2. Policy, procedure, and process actions (27.3%; n = 6) – Standardization of clinical processes and procedures 3. Equipment actions (18.2%; n = 4) – New equipment, upgrade equipment, or remove equipment 1. Training and education actions (27.9%; n = 12) – General development of training and educational programs and materials – Development of cognitive aids, signs, checklists 2. Equipment actions (25.6%; n = 11) – New equipment, upgrade equipment, or remove equipment 3. Policy, procedure, and process actions (23.3%; n = 10) – Standardization of clinical processes and procedures – Review and/or develop protocols/processes – Create new policy
Actions by Type of Adverse Event
Action Effectiveness
Actions were evaluated for the most common types of events (Table 5). Consistent across events was the recommendation for standardization of clinical processes and procedures. Simulation-based training actions were recommended for delay in care events and procedure errors. Actions involving developing cognitive aids, signs, and checklists were recommended for delay in care events, medication errors, and equipment failures.
Of the 277 recommended actions, 270 (97.5%) were implemented and had a reported effectiveness (Figure 4). Of the 270 implemented actions, 47.4% had a reported effectiveness of “much better,” 37.0% had a reported effectiveness of “better,” 7.4% had a reported effectiveness of “same,” and the remaining 8.2% were either “not reported” or “not measured.” Reported effectiveness was also evaluated for each action category (Table 6). Within each action category, a majority of the
ARTICLE IN PRESS Volume ■■, No. ■■, ■■ 2017
7
Recommended Actions for Reported ICU Adverse Events for CY 2013 and CY 2014 (N = 277)
Figure 3: Of the 277 recommended actions identified, policy, procedure, and process actions (34.4%) and training and education actions (31.4%) were the most common. CY, calendar year.
Reported Effectiveness of Implemented Actions (N = 270)
Figure 4: Of the 277 recommended actions, 270 (97.5%) were implemented and had a reported effectiveness.
actions implemented had a reported effectiveness of “much better” or “better.” DISCUSSION
The goal of our study was to identify common underlying factors that are involved across the different categories of reported adverse events. The study is unique in that it is the first to systematically explore reported ICU adverse events using a national RCA database, enabling us to examine the
root causes identified for reported adverse events, as well as recommendations for improvements. We were able to identify root causes that were repeatedly identified across multiple categories of adverse events. The types of adverse events we identified in the RCA database are consistent with the types of adverse events that have been identified in previous studies.4,6–8 A majority of the events had either an actual or potential SAC score of 3. This is not surprising, given that events with a SAC score
ARTICLE IN PRESS 8
Gregory S. Corwin, MPH, et al
RCA of ICU Adverse Events
Table 6. Reported Action Effectiveness by Action Categories Reported Effectiveness (%) Action Category Training and Education Actions (N = 87) Policy, Procedure, and Process Actions (N = 94)† Communication Actions (N = 13) Equipment Actions (N = 34)† Documentation Actions (N = 31) Staffing Actions (N = 9) Other (N = 2)
Much Better/ Better Same Worse Other* 85.06
4.60
0
10.34
81.91
9.57
0
8.51
84.62
7.69
0
7.69
82.35
11.76
0
5.88
96.77
3.23
0
0
66.67 100
11.11 0
0 0
22.22 0
*Not Reported; Not Measured. † Because of rounding, adds up to only 99.99%.
of 3 require an RCA and also represent events that have a higher severity and probability of reoccurrence. Most of the reported adverse events identified more than one root cause. This observation is consistent with Reason’s model of organizational safety suggesting that, rather than a single factor, multiple factors have to align for an adverse event to occur.20 However, the relative contribution of any individual factor is difficult to ascertain. Across most of the reported adverse events, a lack of standardized processes and/or protocols was identified as a root cause. Similarly, standardization of clinical processes and procedures was a common recommended action. These observations are consistent with the recent emphasis on the importance of standardized critical care protocols.21 For example, care bundles specifically directed toward a disease process of care have been successful in improving clinical outcomes in the critically ill.22 Similarly, in a prospective study, the implementation of a resident handoff bundle resulted in a significant decrease in medical errors and preventable adverse events in hospitalized children.23 It has been demonstrated that just having protocols does not necessarily result in the protocols being followed or improved outcomes.24 Additional factors such as protocol implementation, compliance, and behavior change may play a role in the success of ICU protocols.24–26 One study of ICU patients reported that just over 50% of eligible patients were receiving care consistent with best-practice treatment guidelines.27 Similarly, in a trauma population, almost 50% of the patients had care with moderate to severe deviation in compliance with best-practice guidelines;28 greater compliance was associated with improved clinical outcomes. Getting physicians to follow protocols and/or guidelines has been shown to be challenging, with multiple barriers
interfering with physician behavior change and adherence with guidelines.29 Effective teamwork, which has been found to be important in overcoming barriers to adhering to protocols and processes of care in the ICU,30,31 involves team communication, team coordination, team decision making, and team leadership.31 A number of interventions have been used to improve team performance.30 For example, implementing a mandatory checklist of protocols and care objectives on rounds was effective in adherence to best practice.32 In an attempt to improve communication and teamwork, the VHA Medical Team Training (MTT) program was developed.33 MTT was based on crew resource management (CRM), with the goal of increasing the understanding of safety threats in the acute care setting.33 The MTT program has been successfully implemented in the VHA system and has been demonstrated to improve patient outcomes and improve teamwork climate.33–35 Similarly, a meta-analysis of 20 studies that examined CRM in acute hospital settings demonstrated a positive response from participants and an increase in their knowledge and behaviors; however, there was insufficient evidence to determine the impact on patient outcomes.36 Almost all the reported adverse events had actions that were intended to address at least one root cause through a training and education action. A variety of training and education actions were recommended, such as simulationbased training; online training modules; updating required competencies; and the addition of cognitive aids, signs, and checklists in the clinical environment. Although training actions are typically identified as weak actions, the literature suggests that simulation-based training may be the best approach.37–42 Simulation-based training has been shown to be effective in improving technical and nontechnical skills and in turn patient safety.43 A study by Paull et al. demonstrated that simulation-based team training in the perioperative setting improves teamwork and communication among staff.41 It should be noted that not all actions are equal. When viewed individually, some actions are considered to be stronger, and other actions weaker. NCPS developed an action hierarchy, with actions categorized in terms of their relative strength.42 Stronger actions are recognized as those that standardize or simplify processes, while weaker actions are recognized as those involving general training or new procedures and policies.42 In looking at adverse drug events, Mills et al. found that actions that focused on bedside clinical care improvements were associated with improved outcomes.44 However, although general training and education have been identified as weaker actions, more immersive forms of training and education such as MTT and simulation-based training have been shown to be effective in improving clinical care.33–35,37,39,41 Individual sites implemented and reported the effectiveness for more than 97% of the recommended actions.
ARTICLE IN PRESS Volume ■■, No. ■■, ■■ 2017
Most of the reported adverse events were addressed by several actions. Overall, 84.4% of the actions that were implemented resulted in a reported effectiveness of either “much better” or “better.” Despite the self-reported nature of this information, the findings do suggest that the actions recommended and subsequently implemented by sites were effective. In addition, consistent across the different action categories, the majority of actions for a given action category had a reported effectiveness of “much better” or “better.” On the basis of our findings, we suggest the following recommendations for improvement of patient safety in the ICU: • Development and implementation of standardized protocols and processes of care, including strategies for education and compliance with protocols • Development and implementation of team training programs designed using CRM concepts • Simulation-based training for improvement of technical and nontechnical skills Limitations
This study had several limitations. The unique patient population and organizational structure of the VHA may limit the generalizability of the findings to medical centers outside the VHA. The primary limitation, however, is the potential for the underreporting of adverse events. One of the reasons there may be underreporting is that the RCA system relies on reporting by frontline personnel. Different providers may have different views on whether a case is in fact an adverse event. For example, one study comparing adverse event reporting rates of ICU staff to that of observers found that the observers’ reporting rate was nearly double that of the ICU staff.45 In addition, because of the complexity of ICU care, adverse events may be missed when they are instead viewed as a complication of care. We also explored whether facility complexity levels between the facilities in which the RCA cases included in our study occurred and the facilities that did not have RCA cases, to find that there were no significant differences. In addition to underreporting, some adverse events that are reported may not be reviewed using an individual RCA. Individual RCAs are mandatory only for events with SAC scores of 3. For events with a SAC score < 3, individual hospitals decide whether or not to perform an RCA. Although hospital resources and priorities could affect this decision, SAC scores are intended to help hospitals identify and direct their resources to serious events even if they are near misses. An adverse event may also be addressed in an aggregate review of several adverse events that occur throughout a given year at a single site. Moreover, RCA is only one tool available for the review of adverse events; events with lower SAC scores may be addressed by using other patient safety tools. Given the reported frequency of ICU adverse events, particularly those that are life-threatening or fatal, there was a relatively
9
low number of reported ICU adverse events that resulted in an RCA. Although it is unclear whether the number of RCAs is the “correct” number, the fact that the types of events captured in the RCAs corresponds to the types of events reported in the literature suggests that the included RCAs captured the significant adverse events occurring in the ICU.4,6,8 Our analysis of a given action’s effectiveness is also limited. Action effectiveness was self-reported by each site, so we could not evaluate how a given site determined the effectiveness of a given action. In addition, although each action’s effectiveness was reported, the assessment of a single action’s effectiveness might have reflected its implementation in combination with several other actions. This may explain why actions that are noted as being weaker are reported to have a “much better” or “better” effectiveness after being implemented. Since health care’s adoption of the RCA process in the 1990s, RCA itself has come under increased scrutiny. Recently, questions have been raised about the general effectiveness of the RCA process in health care, and recommendations have been suggested to improve it.46–49 Although the VA RCA process provides valuable information on adverse events, there are areas for improvement, particularly regarding follow-up (implemented actions and the sharing of lessons learned). Most important, RCA is most effective when used as part of a set of tools, such as failure mode and effects analysis (FMEA), which have received much less attention in health care—and are used in the VA system, although to a lesser degree.47,50 Efforts are under way to make the FMEA process easier and more accessible to VA medical centers, as well as to make the RCA process more widely available with results that are more “SMART” (specific, measurable, achievable, relevant, time-sensitive). Finally, the importance of sharing lessons learned across VA medical centers, which is a key priority, cannot be overemphasized. CONCLUSION
Reported ICU adverse events typically had several root causes, with protocol and process-of-care issues (lack of standardized approaches) consistently identified as problems. Actions often included standardization of processes and the development and implementation of appropriate training and education. Acknowledgments. This material is the result of work supported with resources and the use of facilities at the US Department of Veterans Affairs (VA) National Center for Patient Safety Field Office in White River Junction, Vermont, and the VA National Center for Patient Safety in Ann Arbor, Michigan. Disclaimer. The views expressed in this article do not necessarily represent the views of the US Department of Veterans Affairs or of the United States Government. Presentations. An abstract of this work was presented at the American Thoracic Society (ATS) International Conference, May 19–24, 2017, Washington, DC. Conflicts of Interest. All authors report no conflicts of interest.
ARTICLE IN PRESS 10
Gregory S. Corwin, MPH, et al
Gregory S. Corwin, MPH, formerly Patient Safety Fellow, US Department of Veterans Affairs (VA) National Center for Patient Safety Field Office, White River Junction, Vermont, is Medical Student, University of Arkansas for Medical Sciences College of Medicine, Little Rock. Peter D. Mills, PhD, MS, is Director, VA National Center for Patient Safety Field Office, White River Junction, Vermont, and Adjunct Associate Professor, Department of Psychiatry, Geisel School of Medicine at Dartmouth, Hanover, New Hampshire. Hasan Shanawani, MD, is Patient Safety Physician, VA National Center for Patient Safety, Ann Arbor, Michigan, and Clinical Instructor, Division of Pulmonary and Critical Care Medicine, University of Michigan, Ann Arbor. Robin R. Hemphill, MD, MPH, is Veterans Health Administration Chief Patient Safety and Risk Awareness Officer, and Director, VA National Center for Patient Safety, Ann Arbor. Please address correspondence to Peter D. Mills, [email protected].
REFERENCES 1. Institute of Medicine. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Press, 2000. 2. James JT. A new, evidence-based estimate of patient harms associated with hospital care. J Patient Saf. 2013;9:122–128. 3. Abramson NS, et al. Adverse occurrences in intensive care units. JAMA. 1980 Oct 3;244:1582–1584. 4. Garrouste-Orgeas M, et al. Selected medical errors in the intensive care unit: results of the IATROREF study: parts I and II. Am J Respir Crit Care Med. 2010 Jan 15;181:134–142. 5. Rothschild JM, et al. The Critical Care Safety Study: the incidence and nature of adverse events and serious medical errors in intensive care. Crit Care Med. 2005;33:1694–1700. 6. Thomas AN, Taylor RJ. Review of patient safety incidents reported from critical care units in North-West England in 2009 and 2010. Anaesthesia. 2012;67:706–713. 7. Valentin A, et al. Patient safety in intensive care: results from the multinational Sentinel Events Evaluation (SEE) study. Intensive Care Med. 2006;32:1591–1598. 8. Welters ID, et al. Major sources of critical incidents in intensive care. Crit Care. 2011;15:R232. 9. Steyrer J, et al. Attitude is everything? The impact of workload, safety climate, and safety tools on medical errors: a study of intensive care units. Health Care Manage Rev. 2013;38:306–316. 10. Merino P, et al. Adverse events in Spanish intensive care units: the SYREC study. Int J Qual Health Care. 2012;24:105– 113. 11. Garry DA, et al. A prospective multicentre observational study of adverse iatrogenic events and substandard care preceding intensive care unit admission (PREVENT). Anaesthesia. 2014;69:137–142. 12. US Department of Veterans Affairs, Veterans Health Administration. VHA National Patient Safety Improvement Handbook. Mar 4, 2011. Accessed Jun 18, 2017. http:// www.va.gov/vhapublications/ ViewPublication.asp?pub_ID=2389. 13. Bagian JP, et al. Developing and deploying a patient safety program in a large health care delivery system: you can’t fix what you don’t know about. Jt Comm J Qual Improv. 2001;27:522–532. 14. Giardina TD, et al. Root cause analysis reports help identify common factors in delayed diagnosis and treatment of outpatients. Health Aff (Millwood). 2013;32:1368–1375. 15. Lee A, et al. Root cause analysis of serious adverse events among older patients in the Veterans Health Administration. Jt Comm J Qual Patient Saf. 2014;40:253–262. 16. Miller KE, et al. Wrong-side thoracentesis: lessons learned from root cause analysis. JAMA Surg. 2014;149:774–779.
RCA of ICU Adverse Events 17. Mills PD, et al. Suicide attempts and completions in Veterans Affairs nursing home care units and long-term care facilities: a review of root-cause analysis reports. Int J Geriatr Psychiatry. 2016;31:518–525. 18. US Department of Veterans Affairs. Veterans Health Administration: About VHA. (Updated: Mar 9, 2017.) Accessed Jun 18, 2017. http://www.va.gov/health/aboutVHA .asp. 19. Salvendy G, editor. Handbook of Human Factors and Ergonomics, 3rd ed. Hoboken, NJ: John Wiley & Sons, 2006. 20. Reason J. Human error: models and management. BMJ. 2000 Mar 18;320:768–770. 21. Weled BJ, et al. Critical care delivery: the importance of process of care and ICU structure to improved outcomes: an update from the American College of Critical Care Medicine Task Force on Models of Critical Care. Crit Care Med. 2015;43:1520–1525. 22. Dellinger RP, et al. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med. 2013;39:165–228. 23. Starmer AJ, et al. Rates of medical errors and preventable adverse events among hospitalized children following implementation of a resident handoff bundle. JAMA. 2013 Dec 4;310:2262–2270. 24. Sevransky JE, et al. Protocols and hospital mortality in critically ill patients: the United States Critical Illness and Injury Trials Group Critical Illness Outcomes Study. Crit Care Med. 2015;43:2076–2084. 25. Bonello RS, et al. An intensive care unit quality improvement collaborative in nine Department of Veterans Affairs hospitals: reducing ventilator-associated pneumonia and catheterrelated bloodstream infection rates. Jt Comm J Qual Patient Saf. 2008;34:639–645. 26. Centofanti JE, et al. Use of a daily goals checklist for morning ICU rounds: a mixed-methods study. Crit Care Med. 2014; 42:1797–1803. 27. Ilan R, et al. Knowledge translation in critical care: factors associated with prescription of commonly recommended best practices for critically ill patients. Crit Care Med. 2007;35:1696–1702. 28. Rice TW, et al. Deviations from evidence-based clinical management guidelines increase mortality in critically injured trauma patients. Crit Care Med. 2012;40:778–786. 29. Cabana MD, et al. Why don’t physicians follow clinical practice guidelines? A framework for improvement. JAMA. 1999 Oct 20;282:1458–1465. 30. Dietz AS, et al. A systematic review of teamwork in the intensive care unit: what do we know about teamwork, team tasks, and improvement strategies? J Crit Care. 2014;29:908– 914. 31. Reader TW, et al. Developing a team performance framework for the intensive care unit. Crit Care Med. 2009; 37:1787–1793. 32. Byrnes MC, et al. Implementation of a mandatory checklist of protocols and objectives improves compliance with a wide range of evidence-based intensive care unit practices. Crit Care Med. 2009;37:2775–2781. 33. Dunn EJ, et al. Medical team training: applying crew resource management in the Veterans Health Administration. Jt Comm J Qual Patient Saf. 2007;33:317–325. 34. Neily J, et al. Association between implementation of a medical team training program and surgical mortality. JAMA. 2010 Oct 20;304:1693–1700. 35. Carney BT, et al. Improving perceptions of teamwork climate with the Veterans Health Administration medical
ARTICLE IN PRESS Volume ■■, No. ■■, ■■ 2017
36.
37.
38.
39.
40.
41.
team training program. Am J Med Qual. 2011;26:480– 484. O’Dea A, O’Connor P, Keogh I. A meta-analysis of the effectiveness of crew resource management training in acute care domains. Postgrad Med J. 2014;90:699– 708. Barsuk JH, et al. Simulation-based mastery learning reduces complications during central venous catheter insertion in a medical intensive care unit. Crit Care Med. 2009;37:2697– 2701. Gold JA, et al. Integrating the electronic health record into high-fidelity interprofessional intensive care unit simulations. J Interprof Care. 2015;29:562–563. Gundrosen S, Solligård E, Aadahl P. Team competence among nurses in an intensive care unit: the feasibility of in situ simulation and assessing non-technical skills. Intensive Crit Care Nurs. 2014;30:312–317. Hebbar KB, et al. Simulation-based paediatric intensive care unit central venous line maintenance bundle training. Intensive Crit Care Nurs. 2015;31:44–50. Paull DE, et al. The effect of simulation-based crew resource management training on measurable teamwork and communication among interprofessional teams caring for postoperative patients. J Contin Educ Nurs. 2013;44:516–524.
11 42. US Department of Veterans Affairs (VA) National Center for Patient Safety. Root Cause Analysis Tools. Washington, DC: VA, 2015. 43. Naik VN, Brien SE. Review article: simulation: a means to address and improve patient safety. Can J Anaesth. 2013;60: 192–200. 44. Mills PD, et al. Effective interventions and implementation strategies to reduce adverse drug events in the Veterans Affairs (VA) system. Qual Saf Health Care. 2008;17:37–46. 45. Capuzzo M, et al. Reporting of unintended events in an intensive care unit: comparison between staff and observer. BMC Emerg Med. 2005 May 27;5:3. 46. Card AJ. The problem with ‘5 whys.’ BMJ Qual Saf. Epub 2016 Sep 2. 47. Peerally MF, et al. The problem with root cause analysis. BMJ Qual Saf. 2017;26:417–422. 48. Wu AW, Lipshutz AK, Pronovost PJ. Effectiveness and efficiency of root cause analysis in medicine. JAMA. 2008 Feb 13;299:685–687. 49. National Patient Safety Foundation (NPSF). RCA2: Improving Root Cause Analyses and Actions to Prevent Harm. Boston: NPSF, 2015. 50. Dixon-Woods M, et al. Safer Clinical Systems: Evaluation Findings. London: The Health Foundation, 2014.