Otolaryngol --head neck surg - 2019 - Balakrishnan - Patient Safety Quality Improvement Primer Part II Prevention of Harm.pdf

SPECIAL FOCUS SECTION: PATIENT SAFETY/QUALITY IMPROVEMENT Invited Article Otolaryngology–...

SPECIAL FOCUS SECTION: PATIENT SAFETY/QUALITY IMPROVEMENT Invited Article Otolaryngology– Head and Neck Surgery Patient Safety/Quality Improvement 2019, Vol. 161(6) 911–921 Ó American Academy of Otolaryngology–Head and Neck Primer, Part II: Prevention of Harm Surgery Foundation 2019 Reprints and permission: Through Root Cause Analysis and Action sagepub.com/journalsPermissions.nav DOI: 10.1177/0194599819878683 (RCA2) http://otojournal.org Karthik Balakrishnan, MD, MPH1, Michael J. Brenner, MD2, John W. Gosbee, MD, MS3, and Cecelia E. Schmalbach, MD, MSc4 Sponsorships or competing interests that may be relevant to content are dis- new to health care, RCA has a long history of effective closed at the end of this article. implementation in industry to facilitate understanding of accidents and error prevention. RCA is an analysis tool that focuses on systems, as opposed to individuals, to identify Abstract underlying vulnerabilities contributing to the likelihood of With increasing emphasis on patient safety/quality improve- errors. Its presumed effectiveness is such that the Joint ment, health care systems are mirroring industry in the Commission requires its application to all sentinel events.3,4 implementation of root cause analysis (RCA) for the identifi- RCA methodology is derived from engineering and other cation and mitigation of errors. RCA uses a team approach nonmedical disciplines.1,5 While many otolaryngologists with emphasis on the system, as opposed to the individual, have a conceptual understanding of RCA, practical experi- to accrue empirical data on what happened and why. While ence in conducting RCAs is often lacking. This knowledge many otolaryngologists have a broad understanding of RCA, gap is important because individuals are less likely to practical experience is often lacking. Part II of this patient engage in change if they do not understand the process. safety/quality improvement primer investigates the manner While a clinical perspective is extremely helpful, clinical in which RCA is utilized in the prevention of medical errors. experience alone is inadequate in the identification of proxi- Attention is given to identifying system errors, recording mal causes of an adverse event and the necessary actions to adverse events, and determining which events warrant RCA. modify a system to prevent recurrence. The primer outlines steps necessary to conduct an effective The aim of part II of this PS/QI primer is to explain the RCA, with emphasis placed on actions that arise from the history of RCA, demonstrate how an RCA is conducted, RCA process through the root cause analysis and action (or and show how the findings of an RCA are used to prevent RCA2) rubric. In addition, the article provides strategies for medical errors. The hallmarks of RCA are rigor, transpar- the implementation of RCA into clinical practice and medi- ency, and unbiased analysis. Emphasis is placed on actions cal education. that arise from the RCA process through the root cause analysis and action (RCA2) rubric.6 In addition, this article Keywords provides opportunities and strategies for implementing RCA into clinical practice and medical education. patient safety, quality improvement, root cause analysis, root cause analysis and action, total quality management, failure mode and effects analysis Received April 18, 2019; accepted September 6, 2019. 1 Mayo Clinic Department of Otorhinolaryngology and Mayo Children’s Center, Rochester, Minnesota, USA 2 Department of Otolaryngology–Head and Neck Surgery, School of Medicine, University of Michigan, Ann Arbor, Michigan, USA D ating back to the Hippocratic oath, health care pro- 3 Departments of Biomedical Engineering, Internal Medicine, and Graduate viders and organizations have embraced the idea of Medical Education, University of Michigan, Ann Arbor, Michigan, USA causing ‘‘zero harm’’ to patients. Leading organiza- 4 Department of Otolaryngology–Head and Neck Surgery, Lewis Katz tions such as the Joint Commission recently renewed the School of Medicine at Temple University, Philadelphia, Pennsylvania, USA focus on this objective.1,2 However, the practical path toward zero harm through patient safety/quality improve- Corresponding Author: Cecelia E. Schmalbach, MD, MSc, Department of Otolaryngology–Head and ment (PS/QI) strategies remains unclear. Root cause analy- Neck Surgery, Lewis Katz School of Medicine at Temple University, 440 N sis (RCA) is an easily understood and practical tool that has Broad St, Kresge W, Third Floor, Philadelphia, PA 19140, USA. taken on central importance in this effort. While relatively Email: [email protected]  10976817, 2019, 6, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599819878683 by Scholarly Information Univ Lib, Wiley Online Library on [31/08/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 912 Otolaryngology–Head and Neck Surgery 161(6) Table 1. Key Principles of TQM7 and Their Health Care Analogs. TQM Principle Health Care Equivalent 1. Customer-focused mentality Patient centered care 2. Empowered employees led by invested leaders, all working for continuous improvement without Culture of safety, just culture fear of retribution 3. Process centered with continuous monitoring of performance measures and resulting data-driven Culture of safety decision making 4. Interconnected organization with integration across disciplines to improve quality Culture of safety 5. Strategic plan has quality as a central component Culture of safety 6. Continuous improvement Continuous improvement Abbreviation: TQM, total quality management. Background and Impact of RCA an adverse event transpires. Correcting these vulnerabilities in a system can decrease the risk of adverse events. RCA began as a key component of total quality management Accordingly, error and adverse event analysis should ide- (TQM). TQM is a business and manufacturing approach to ally integrate an understanding of individuals’ decision pro- quality improvement that emphasizes a customer-centered cesses and the system in which those processes occur.11 organization and the involvement of every member of the One example has been developed by the United Kingdom’s organization in constant improvement of quality. Individuals National Health Service; this example forms the first part of interested in delivering quality health care will likely find the flowchart in Figure 1. A critical point in this workflow this concept closely aligned with the goal of providing high- is the ‘‘substitution test,’’ which asks whether a provider quality patient-centered care. TQM has several key tenets (with similar training and expertise) would make the same that are as applicable to excellent health care organizations as decision or error under the same circumstances in the same to manufacturing (Table 1).7 system.9 An answer of ‘‘yes’’ suggests that system factors As with many other innovations from manufacturing, the were important contributors to the error in question. application of TQM-like strategies to health care, including To operationalize this comprehensive approach to error RCA, has led to important improvements in safety in quality. analysis, 3 questions must be answered: For example, the Institute for Healthcare Improvement sug- gests that a basic set of quality initiatives adopted by 3100 Question 1: How will adverse events be detected and hospitals over 2 years led to the prevention of.122,000 recorded? ‘‘needless deaths.’’8 The systematization of PS/QI in medicine Question 2: Which adverse events warrant RCA? has fundamentally changed the culture of health care. The Question 3: How does RCA fit into a quality and pro- result is a model that centers on honest appraisal of the com- cess improvement program? plexity of health care delivery and its attendant vulnerabilities; RCA operationalizes that model to facilitate actual improve- ment. As stated by one author, RCA ‘‘induces thoughtful con- Detection and Recording of Adverse Events sideration of the causal complexities of classes of medical The cornerstone of event detection has long been voluntary adverse events. If used with care and intelligence it can reveal reporting, whether through departmental activities such as mor- potential hazards and instruct ways to mitigate them.’’9 bidity and mortality (M&M) conferences,12 institutional regis- tries,13 or national databases such as the American Academy of Identifying System Failure Otolaryngology—Head and Neck Surgery Foundation event Adverse events and near misses have a complex relationship reporting system.14 While necessary for event detection and as with human error. Medical errors that cause harm involve a marker for institutional safety culture,11 voluntary reporting is individuals’ perception and interpretation of data in the con- insufficient for any comprehensive quality program due to text of their work. Underlying systems and processes influ- selective capture of events, inaccurate event recall, availability ence the quality, transmission, and interpretation of those bias,15 inaccurate or poorly detailed reporting, and cumbersome data, as well as the implementation of actions based on data reporting systems.16 Furthermore, clinicians may not report (or interpretation. These defense layers can be thought of as the even misreport) events that might cast them in a negative light, slices of Swiss cheese in James Reason’s classic model of particularly at institutions where a ‘‘name/blame/shame’’ culture adverse event causation.10 This model emphasizes a com- exists. plex system in which a series of barriers or defenses exist to Voluntary reporting should be supplemented by auto- prevent hazards, but each barrier has weaknesses, or holes, mated or semiautomated event detection processes that cap- like Swiss cheese. When all barriers fail and the holes align, ture events reaching the patient and near misses.17 Ideally,  10976817, 2019, 6, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599819878683 by Scholarly Information Univ Lib, Wiley Online Library on [31/08/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Balakrishnan et al 913 Figure 1. Overview of event analysis. The analysis of an adverse event should incorporate both system- and individual-level considerations. One example of an analytic pathway is shown here. Dark gray boxes indicate relevance to individual judgment or bias concerns; lighter boxes indicate other considerations. HFACS, Human Factors Analysis and Classification System. this capture process and voluntary reporting produce results detailed RCA and which might be better addressed through that are legally protected. Otolaryngology-specific reporting more expeditious but less rigorous review strategies. structures are rare and limited to focused questions, such as RCA is resource intensive; therefore, only a minority of tracking posttonsillectomy hemorrhages.18 However, authors adverse events can be subjected to formal analysis. As such, have proposed a variety of approaches in the broader medi- the impact of a quality program is predicated on effective cal literature. Many of these strategies rely on automated case selection. Some institutions apply RCA to all occur- searches of the electronic health record,19 natural language rences from a predetermined list of events or case types. For processing of electronic record data,20 or statistical methods example, the American College of Surgeons’ Children’s of testing and validation.21 Others rely on human data entry Surgery Verification Program lists 36 events that ‘‘should’’ mandated by a ‘‘hard stop’’ in the charting process22; the or ‘‘must’’ be reviewed in surgical patients aged \18 years, recorded data are flagged or sent to a database for subse- ranging from dental trauma to death within 30 days.31 Other quent analysis as appropriate.23 organizations triage events based on harm level, using Regardless of the specific strategy used to identify and widely accepted scales such as the Agency for Healthcare capture events, standardized definitions must be developed Research and Quality’s Common Format or Clavien- for tracked events.24 These definitions may be operationa- Dindo12 or similar homegrown harm rating scales. lized through actual clinical tools25,26 or simply dissemi- These strategies have a shared weakness: the inattention nated and used consistently in the medical record by all to potentially serious problems, often referred to as ‘‘near individuals entering and analyzing event-related data in a misses,’’ that do not lead to direct patient harm. A more given system. Standardized definitions are also helpful in thorough approach is ‘‘risk-based prioritization,’’6 which developing trigger tools,27 which are data values or terms gives the most attention to problems that have the highest that are typically associated with adverse events. For exam- risk of causing patient harm, regardless of whether that ple, the use of medicinal leeches likely would indicate harm actually occurred. This prioritization is based on tissue or flaps with vascular compromise. These tools can potential degree of harm and probability of the event. These be statistically validated.28 The more specific a trigger tool vulnerabilities can be identified through methods such as is for a given adverse event, the better. Other strategies failure mode and effects analysis (FMEA), discussed later. have used human factors–related trigger tools, such as auto- One risk-based prioritization tool is the Safety Assessment mated radiofrequency identification tracking29 or biometric Code matrix developed by the Veterans Affairs’ National and physiologic data, including heart rate variability30 Center for Patient Safety.6 (reflecting workload), to collect data on processes and work- The Safety Assessment Code rates harm severity as cata- flows relevant to error analysis. strophic, major, moderate, or minor. Each category has clear definitions for harm to patients, visitors, and staff. Similarly, Which Events Warrant RCA? event probability is categorized and defined as frequent, The event capture strategies described here, if implemented occasional, uncommon, or remote. Each combination of effectively, will capture large numbers of events. The next severity and probability is assigned a score of 1 through 3, challenge is to determine which events are ‘‘worthy’’ of a with events scoring 3 warranting RCA (Table 2). This  10976817, 2019, 6, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599819878683 by Scholarly Information Univ Lib, Wiley Online Library on [31/08/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 914 Otolaryngology–Head and Neck Surgery 161(6) Table 2. National Patient Safety Foundation Safety Assessment contributing factors (RCCFs), defining corrective actions, Code Matrix.59,a measuring outcomes, and communicating results (Table Severity 3).32 As outlined here, success is predicated on reporting of errors that results in an unbiased, systematic analysis. Probability Catastrophic Major Moderate Minor The ideal RCA team comprises 4 to 6 individuals dedi- cated to focusing on systems issues associated with the Frequent 3 3 2 1 adverse event.1 Diversity is important, and participants Occasional 3 2 1 1 should include doctors, allied health professionals, adminis- Uncommon 3 2 1 1 trators, and quality improvement experts. In an effort to Remote 3 2 1 1 minimize bias, individuals involved in the case should be a Reprinted from www.IHI.org with permission of the Institute for excluded.33,34 Healthcare Improvement, Ó2018. The initial flow diagram (Figure 2) is produced to show the chain of events preceding the incident.35 Trigger ques- tions are used to identify areas of inquiry that may other- rating system can be applied to actual events and to near wise be overlooked.36,37 Individuals working in areas misses. identified and outlined on the flow diagram are interviewed to elucidate specifics related to the surroundings, events that Performing RCA transpired, and the potential ‘‘why.’’ It is important to note RCA steps include identifying cases, organizing a team, that errors are divided into 2 broad categories. Active errors generating case-specific instruments (flow diagram, event are unsafe actions directly linked to the adverse event/harm, story map, and cause-effect diagram), identifying root cause such as pilot error in an airline crash. Latent errors are Table 3. Process of RCA.a Step in Process Components of Step 1. Identify adverse event  Honest and open reporting of adverse events  Committee review of clinical documentation to understand basics of the following: What event happened? When? Who was involved? How and why did it happen?  Identify appropriate RCA investigations 2. Organize a team  Team should consist of 4-6 members of clinicians, supervisors, and quality improvement experts with fundamental knowledge of specific area of interest  Ensure that, despite members having different levels of authority, everyone be treated as an equal  Members should not be directly involved with the case in question  Appoint an unbiased team leader/facilitator 3. Develop an initial flow diagram  Use a flowchart to describe the processes leading to the event  Organize the information to reach a mutual understanding of the problem 4. Develop an event story map  Use triggering questions to guide further investigation  Conduct thorough interviews with all parties involved in event  Thoroughly review clinical documentation surrounding the event 5. Develop a cause-and-effect diagram  Identify a single problem statement  Identify actions and conditions that caused the problem statement  These categories should address communication problems, policies, rules, procedures, and human errors leading to the event 6. Identify root cause contributing factors  Describe how a cause led to an effect and increased the likelihood of adverse event  Apply 5 rules of causation for crafting root cause contributing factor statements 7. Develop corrective actions  Identify barriers and risk reduction strategies to prevent root cause from recurring  Multiple actions may be required  Implement a trial test of corrective action 8. Measure outcomes  Develop outcome measurements to ensure appropriate implementation of actions  Track quantifiable data to document effectiveness of actions over time  Evaluate and fine-tune improvement efforts if needed 9. Communicate results (loop closure)  Communicate results of RCA to all staff involved in event and more broadly if applicable Abbreviation: RCA, root cause analysis. a Adapted from Charles et al32 (http://creativecommons.org/licenses/by/4.0/).  10976817, 2019, 6, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599819878683 by Scholarly Information Univ Lib, Wiley Online Library on [31/08/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Balakrishnan et al 915 Paent with Paent is Orders placed Suconing and tracheostomy admied to by resident for Humidificaon presents to ED short stay floor tracheostomy equipment not with dyspnea for evaluaon care at bedside Rapid Response Difficules with Floor staff cite Residents note called due to trach care and concerns with no acon with hypoxia /airway educaon seen care orders for orders, crusted compromise on am rounds tracheostomy tracheostomy Figure 2. The initial flow diagram is a chronological timeline showing the chain of events preceding the adverse event—in this case, a tra- cheostomy occlusion on a floor without otolaryngology expertise. Trigger questions will help identify areas for further inquiry. ED, emer- gency department. Efforts to find necessary supplies Tracheostomy occludes, Paent is admied to a on the floor prove unsuccessful paent experiences hypoxia general “Short Stay” floor (Contribung Event) (Adverse Event & Time) (Early Event) Are supplies stocked on floor? How was this floor locaon chosen? Do these supplies expire? Was the hospital short on beds? Issues of space/ inventory? What pressures were affecng the ED? Paent Safety Report, Sucon and humidificaon orders Floor manager review Rapid Response called to are not acted upon across shis (Event Follow-up) resuscitate the paent (Specific Precipitang Event) (Event Discovery) Were the orders clear and viewable? Was the RN/RT staffing adequate? Do policies differ by hospital floor? Figure 3. The event story map allows the root cause analysis team to identify and view significant events that occurred before and after the adverse event. Chart reviews and personnel interviews can help answer questions and clarify conditions that may have contributed to the adverse outcome. This example relates to the same tracheostomy ‘‘near miss’’ case. ED, emergency department; RN, registered nurse; RT, respiratory therapist. Adapted from Charles et al32 (http://creativecommons.org/licenses/by/4.0/). contributing factors that are not immediately associated with to link a problem statement to conditions and actions. an adverse event/harm but have the potential to become a Whereas the fishbone diagram is widely used as an intuitive contributing factor. For example, organizational restructur- and visually powerful tool, the cause-and-effect diagram has ing that reduces pilot training hours may eventually lead to the advantage of allowing deeper analysis of successive an error. causal layers. Documents and records providing additional information Ultimately, the information incorporated into the cause- to include hospital procedures, guidelines, equipment, and and-effect diagram or fishbone diagram enables the team to other related data are reviewed. In addition, the site where explore contributing factors and define an RCCF state- the event occurred is visited. Equipment, accessories, and ment.39 The structure of an RCCF statement is ‘‘[Cause x] other tools, such as the Food and Drug Administration’s led to [effect y], predisposing to [event z].’’32 The RCCF MAUDE database (Manufacturer and User Facility Device statement is critical, as it dictates implementation of correc- Experience), are examined and researched.38 MAUDE con- tive outcomes and how change will be measured. After the tains medical device reports submitted to the Food and RCCF statement is drafted, the 5 rules of causation (Box 1) Drug Administration by mandatory reporters (manufactur- are used to refine the statement.37,40 These rules include (1) ers, importers, and device user facilities) and voluntary clearly showing the cause-effect relationship, (2) using pre- reporters (patients, consumers, and health professionals). cise descriptors to explain what occurred, (3) providing a This information is incorporated into an event story map preceding cause for any human error, (4) identifying a pre- (Figure 3), which captures details of the inciting events, ceding cause in instances where procedure was not fol- adverse event, other contributory events, discovery of the lowed, and (5) determining whether there was a duty to act. adverse event, and follow-up. A fishbone diagram (Figure For example, a surgeon might be inclined to hold a nurse 4) or cause-and-effect diagram (Figure 5) is then developed responsible for delayed detection of free flap compromise;  10976817, 2019, 6, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599819878683 by Scholarly Information Univ Lib, Wiley Online Library on [31/08/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 916 Otolaryngology–Head and Neck Surgery 161(6) Respiratory Using personnel Personnel RT in a Etc. Therapist (RT) at maximum level workload not hurry with uses wrong med of effort has rounely many tasks (Acon) become roune reviewed Different Lowest bid for a certain Etc. pharmaceucal medicaon in that companies used fiscal year Label located in different areas of med No FDA Few error reports sent to Paent given container Etc. regulaons about FDA about wrong RT med nebulized Epinephrine standard locaon administraon epinephrine container looks instead of like saline (Condion) saline, suffers Embossed plasc Ink on this type Most economical Etc. cardiac event label low of plasc would plasc for this contrast leach through formulaon Medicaon label Lighng level RT medicaon Prior input on difficult to read lower than storage area No complaints work place Etc. needed for from RT about (Condion) with low lumen condion was reading small lighng lightbulbs ignored font Figure 4. The fishbone diagram provides a visual depiction of the contributory factors and underlying causes leading to an adverse event— in this case, highlighting issues of communication, equipment/supplies, policy, and ordering procedures. PRN, as needed; RN, registered nurse; RT, respiratory therapist. however, under rules of causation, a nurse has a duty to act various actions are then placed on the diagram. Priority is (in this case, alert the surgeon to flap compromise) only if then given to the actions identified as having the highest she or he has been informed of the duty to perform flap sur- probability of corrective outcome at the lowest cost in time veillance and to notify the surgeon of a change. and resources. An important consideration is that preventing future harm requires not only addressing the points of fail- Box 1: Five Rules of Causation ure in the index case but also correcting latent systems defects that have potential for similar harm. 1. Clearly show the cause-and-effect relationship. When changes are being implemented, it is important to 2. Use specific and accurate descriptors for what occurred, rather use outcome measures that are specific, quantifiable, and than negative or vague words. time limited so that effect of the system change can be veri- 3. Human errors must have a preceding cause. fied.5 Finally, through ‘‘loop closure’’ strategies described 4. Violations of procedure are not root causes but must have a in the next section, the results should be communicated to preceding cause. all individuals involved in the event and more broadly if 5. Failure to act is only causal when there is a preexisting duty to appropriate.5 Improving capabilities in RCA can be act. achieved by educating residents and the faculty in these pro- Adapted from Charles et al32 (http://creativecommons.org/licenses/ cedures, including hands-on exposure through experiential by/4.0/). learning.43-46 RCCF statements are placed immediately preceding the Integrating RCA into Broader Quality and location of a known system vulnerability; several such state- Process Improvement Programs ments may be placed, and more than one correction may be As mentioned earlier, properly functioning event detection necessary.41 In this manner, the holes in the Swiss cheese systems will capture many more cases than warrant com- model detailed at the beginning of this primer are addressed plete RCA. Events not reaching the threshold to undergo and closed.10 RCA may be appropriate for expedited review or aggregated The goal of corrective actions is to implement specific review, in which events are analyzed in batches to identify changes that will lead to sustained system improvement. As shared factors or solutions.6 These reviews may still lead to resources for implementation of changes are finite, the RCA specific actions and follow-up plans as might an RCA. team must work collaboratively to identify the most impact- Regardless of the method of analysis—RCA or otherwise— ful changes to address the root contributing causes. Pareto event analysis has little value without subsequent corrective analysis can be applied when there are several possible actions and follow-up, as described in the previous section. courses of action competing for consideration.42 This As a result, many quality programs refer to RCA2. When approach utilizes a diagram with cost on the x-axis and the root causes are identified through RCA of an event, they expected benefit of an action on the y-axis. Sticky notes for must be addressed to prevent similar events or to mitigate  10976817, 2019, 6, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599819878683 by Scholarly Information Univ Lib, Wiley Online Library on [31/08/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Balakrishnan et al 917 Paent Factors Process/Procedures People Aphonic aer No order set for Lack of staff tracheostomy humidificaon comfort/experse Poor manual The “Sucon PRN” Inconsistent dexterity order is ambiguous paent educaon Tracheostomy Occlusion with Lack of role clarity Anoxic Injury Extra inner cannulas Variable staffing of for RT vs RN dues not available on floor allied professionals Expectaons vary by unit Materials Environment Policies Figure 5. This sample cause-and-effect diagram depicts an adverse drug event inspired by real events reported to the ISMP (Institute for Safe Medication Practice). The problem is linked to proximal causes, including one action and many conditions. The cause-and-effect dia- gram is read from left to right, connected by ‘‘caused by’’ statements (arrows) leading to successively deeper root causes of the adverse event. Dashed arrows marked ‘‘Etc’’ reflect that further inquiry is required. FDA, Food and Drug Administration; RT, respiratory therapist. harm from future such events. Actions to address these root event review strategy for M&M. Other programs have causes can be classified by the action hierarchy6 into stron- incorporated RCA into a separate patient safety forum to ger, intermediate, or weaker actions. The National Patient create a more neutral setting for discussion. In either case, Safety Foundation recommends that all events undergoing effective application of RCA techniques in any formal case RCA2 end with at least one stronger or intermediate action, review process requires that the process have a clear struc- though weaker actions may be necessary to lay the ground- ture with the following features12,50: work for implementation of these stronger actions. Once actions have been identified and pursued, loop clo-  Mandatory attendance sure and long-term follow-up are essential. Loop closure,  Respected clinical leader as moderator drawn from closed-loop manufacturing systems,47 refers to  Clear trigger criteria for reporting and case selection the dissemination of RCA2 conclusions to key stakeholders (may use risk-based prioritization for case selection) who might encounter similar situations in the future, includ-  Appropriate culture of safety ing those working in other parts of the institution. For  Consistent conference format example, analysis of a difficult airway management error  Consistent case presentation template, including might lead to the strong action of switching to a better harm classification (presentation can use RCA struc- videolaryngoscope system. This switch might start in the ture as a template) operating room with anesthesiology but ideally should  Clear identification of actions resulting from case spread to any group or setting managing difficult airways to reviews and individuals responsible for those actions include the intensive care units, the emergency department, (actions can be prioritized with the action hierarchy and postoperative recovery areas. as strong, intermediate, or weaker) Without this essential loop closure, mitigation of harm will be limited to one specific area, while the probability of The exact strategy by which actions resulting from case similar events and associated risk/vulnerability will remain reviews can be incorporated into clinical practice and how the unchanged in the remaining areas. Long-term follow-up, or results of actions are followed up will vary among institutions. control, is necessary to maintain the improvements achieved through RCA2 actions,48 statistical process control, and/or other maintenance and monitoring strategies. Apart from the RCA2 Curriculum in Graduate Medical system changes identified through RCA or RCA2, other Education important considerations are feedback to the patient and Otolaryngologists with RCA expertise bring value to their family of any new knowledge or improvements identified department as well as their institution. RCA capabilities can and support for clinical providers involved in the adverse be enhanced by educating residents, fellows, ancillary staff, event—the ‘‘second victims.’’49 and faculty in these procedures, including hands-on expo- One common venue in which event analysis, action iden- sure through experiential learning.43-46 As front-line health tification, and loop closure elements are combined is the care providers in academic health care with nurses and M&M conference.12 Many departments and institutions uti- other bedside providers, trainees provide a unique perspec- lize M&M as the main driver for identifying and addressing tive on system and process issues leading to adverse events; quality gaps and adverse events. Some programs have taken attending providers may be shielded from and unaware of the additional step of formally using RCA or RCA2 as the many of these issues. However, trainees may be intimidated  10976817, 2019, 6, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599819878683 by Scholarly Information Univ Lib, Wiley Online Library on [31/08/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 918 Otolaryngology–Head and Neck Surgery 161(6) by the prospect of playing a key role on an RCA team. For RCA than that described in this tutorial, the principle stands this reason, a dedicated RCA curriculum is invaluable to that RCA can be used in any clinical setting. The same key trainees, may help improve overall quality of care, and pro- principles of event detection, case selection, structured vides a basis for quality improvement projects for all review, loop closure, and follow-up apply. involved.51 A leading example is the University of Michigan’s RCA2 Reasons Why RCA Is Not Always Effective curriculum for residents, which consists of independent While RCA has become a standard tool for quality improve- study, class sessions, and experiential learning.35,52 This ment programs across all medical and surgical specialties, it mentored curriculum is built on a combination of fundamen- is not infallible. One major problem is the neglect of actions tal concepts, generic examples, and teaching cases devel- once root causes are identified. A recent study of.300 oped by RCA experts based on 2 actual events. Prior to RCAs found that weaker actions, such as training and policy the first 2-hour session, residents complete specialized mod- reinforcement, were most commonly proposed, leading to ules relating to RCA. The resident class identifies an appro- repeated occurrences of similar events.54 In addition, 8 key priate case and reads the background and summary of the failings may threaten RCA effectiveness in health care55: adverse event. Learners create an initial flow diagram so that the first meeting is a working session. At the second 1. The focus on a few specific ‘‘root causes’’ leads to session, the event story map is vetted, and the development linear cause-and-effect thinking and neglect of the of cause-effect diagrams begins by using information complex web of actual and latent contributors that obtained from review of data, interviews, and site visit. In truly lead to events. the third and final session, RCCF statements are refined, 2. Quality of RCAs is highly variable due to lack of corrective actions developed, and plans made to measure training, limited standardization, challenges with outcomes and communicate results. This curriculum has recall of event detail, and variable information been rolled out in 9 programs, including otolaryngology, quality. urology, psychiatry, and radiology. Select RCA actions are 3. Bias, lack of independence, and strict timelines implemented as quality improvement projects. prevent thorough analysis. A recent RCA case from this curriculum is instructive. 4. Unclear effectiveness of, and lack of follow-up for, The case, depicted in the fishbone diagram shown in actions proposed from the RCA. Figure 5, involved a patient with tracheostomy being 5. Lack of loop closure. treated on a nonotolaryngology floor. A near miss occurred 6. A siloed focus on individual events within individ- when the patient failed to receive humidification and devel- ual organizations. oped progressive desaturation from crusting of dried secre- 7. Lack of focus on how the interaction between indi- tions. Tracheostomy tube occlusion resulted in a threatened viduals and systems leads to events. airway. In this case, RCA led to a series of interventions, 8. No single individual is given (or takes) responsibil- including creation of a standardized tracheostomy order set, ity for an event or for corrective actions to remedy initiation of a best practice advisory alert through the elec- that event. tronic health record, and enhanced role clarity for allied health professionals around tracheostomy care. Coincident with these developments, the institution invested in building FMEA: A Preemptive Approach to Harm a difficult airway navigator within the electronic record that Prevention is a standardized method for evaluating and recording By its nature, RCA is a reactive strategy—a trigger event airway considerations, emergency action plans, and signage and associated harm must occur for RCA to have material on patient-specific anatomic considerations. to analyze. Used in isolation, RCA can therefore produce only gradual, evolutionary improvement as various events RCA in Nonacademic Settings occur and are analyzed and as the resulting actions are The previous sections discuss many elements that are implemented to create an organically developed ecosystem common in academic health care settings (trainees, M&M of quality improvement actions. conference, etc). However, RCA is not restricted to acade- The converse of RCA is FMEA, which is a proactive mia; indeed, as mentioned earlier, it did not originate in approach of predicting the consequences of potential system health care at all. Data on best practices for RCA or RCA2 failures and acting to preempt these failures and the associ- implementation in private practice, community health care, ated consequences. This strategy involves asking 3 ques- or rural settings are nonexistent, but the importance of these tions: ‘‘What could go wrong?’’‘‘Why might it go wrong?’’ approaches is nevertheless recognized by leading health and ‘‘What are the consequences?’’56 These questions are care quality groups. For example, the Agency for Healthcare familiar to many physicians and surgeons who ask them- Research and Quality created a 4.5-hour RCA curriculum for selves similar questions when planning treatment regimens primary care practices that may be broadly applicable.53 or operative procedures, when obtaining informed consent, While that curriculum uses a somewhat looser approach to and when weighing different treatment options with patients.  10976817, 2019, 6, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599819878683 by Scholarly Information Univ Lib, Wiley Online Library on [31/08/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Balakrishnan et al 919 Some versions of FMEA also incorporate probability ratings Academy of Otolaryngology—Head and Neck Surgery Foundation and measures of detectability for various failures, in effect Coordinator of Research & Quality/Board of Directors. adding risk-based prioritization as in RCA.57 Sponsorships: None. As has been pointed out, ‘‘if FMEA could do exactly Funding source: None. what it is claimed to be, there would be no need for RCA.’’9 In real-world application, it is unlikely that every References potential failure or vulnerability can be preidentified with subsequent systems developed to prevent all possible 1. Rooney JJV, Lee N. Root cause analysis for beginners. Qual adverse consequences of these vulnerabilities. A more rea- Prog. 2004;37:45-53. listic approach is to implement FMEA and RCA as comple- 2. Joint Commission Center for Transforming Healthcare. High mentary strategies that, if used correctly and in concert, will reliability in health care is possible. https://www.centerfor reduce adverse events and allow effective responses.9 transforminghealthcare.org/high-reliability-in-health-care. Accessed FMEA and RCA both focus on identifying and prioritizing March 4, 2019. high-impact quality gaps so that quality improvement efforts 3. Agency for Healthcare Research and Quality Patient Safety will be optimized for the greatest impact possible. Network. Root cause analysis. https://psnet.ahrq.gov/primers/ primer/10/Root-Cause-Analysis. Accessed March 4, 2019. Conclusion 4. Joint Commission. Sentinel events. https://www.jointcommis sion.org/assets/1/6/CAMH_2012_Update2_24_SE.pdf. Accessed With increasing emphasis on PS/QI, health care systems May 20, 2019. mirror industry in their implementation of RCA for the 5. Carroll JS, Rudolph JW, Hatakenaka S. Lessons learned from identification and mitigation of errors. Successful imple- non-medical industries: root cause analysis as culture change mentation of RCA and RCA2 requires a team approach in at a chemical plant. Qual Saf Health Care. 2002;11:266-269. conducting a methodical, stepwise review. Empirical data 6. National Patient Safety Foundation. RCA2 improving root on what transpired prior to and during the sentinel event are cause analyses and actions to prevent harm. https://cdn.ymaws. explored to identify all potential contributing factors; it is com/www.npsf.org/resource/resmgr/PDF/RCA2_first-online- important to note that there is often more than one reason pub_061615.pdf. Accessed May 20, 2019. why an adverse event occurred. The identification and prior- 7. ASQ. What is total quality management (TQM)? https://asq itization of actions stemming from RCA are essential to pre-.org/quality-resources/total-quality-management. Accessed March vent future events. RCA also provides an opportunity to 4, 2019. capture close calls, near misses, and less serious adverse 8. Baehrend J. 100,000 Lives Campaign: ten years later. Institute events that lead to patient harm. Meanwhile, FMEA offers for Healthcare Improvement website. http://www.ihi.org/com the opportunity to anticipate problems before they occur, munities/blogs/_layouts/15/ihi/community/blog/itemview.aspx? acting as the proactive counterpart to RCA. Ultimately, List=7d1126ec-8f63-4a3b-9926-c44ea3036813&ID=268. Accessed active participation in RCA is an integral step toward creat- March 4, 2019. ing a culture of safety58 and ‘‘zero harm’’ and should be 9. Senders JW. FMEA and RCA: the mantras of modern risk implemented in one’s practice as well as otolaryngology management. Qual Saf Health Care. 2004;13:249-250. training programs. 10. Reason J. Human error: models and management. BMJ. 2000; 320:768-770. Author Contributions 11. Boysen PG. Just Culture: a foundation for balanced account- Karthik Balakrishnan, concept/design/acquisition of information; ability and patient safety. Ochsner J. 2013;13:400-406. draft; revision; final draft approval; accountability; Michael J. 12. Balakrishnan K, Arjmand EM, Nussenbaum B, Snyderman C. Brenner, concept/design/acquisition of information; draft; revi- sion; final draft approval; accountability; John W. Gosbee, con- National multispecialty survey results: comparing morbidity cept/design/acquisition of information; draft; revision; final draft and mortality conference practices within and outside otolaryn- approval; accountability; Cecelia E. Schmalbach, concept/design/ gology. Otolaryngol Head Neck Surg. 2018;158:273-279. acquisition of information; draft; revision; final draft approval; 13. Crandall KM, Almuhanna A, Cady R, et al. 10,000 Good accountability. catches: increasing safety event reporting in a pediatric health care system. Pediatr Qual Saf. 2018;3:e072. Disclosures 14. Vila PM, Lewis S, Cunningham G, et al. A novel patient safety Competing interests: Karthik Balakrishnan, American Academy event reporting tool in otolaryngology. Otolaryngol Head Neck of Otolaryngology—Head and Neck Surgery Foundation Patient Surg. 2017;157:117-122. Safety/Quality Improvement Committee; Springer, Inc: royalties as 15. Balakrishnan K, Arjmand EM. The impact of cognitive and impli- textbook editor; Rosivo, Inc: research grant; Beryl Institute: research grant; Mayo Clinic: research grant; Michael J. Brenner, cit bias on patient safety and quality. Otolaryngol Clin North Am. American Academy of Otolaryngology—Head and Neck Surgery 2019;52:35-46. Foundation Patient Safety/Quality Improvement Committee; 16. Tighe CM, Woloshnyowych M, Brown R, Wears B, Vincent Cecelia E. Schmalbach, teaching honorarium: AO North America C. Incident reporting in one UK accident and emergency CMF Trauma (nonprofit); Oto-HNS Editorial Board; American department. Accid Emerg Nurs. 2006;14:27-37.  10976817, 2019, 6, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599819878683 by Scholarly Information Univ Lib, Wiley Online Library on [31/08/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 920 Otolaryngology–Head and Neck Surgery 161(6) 17. Thomas MK, McDonald RJ, Foley EF, Weber SM. Educational quality%20programs/csv/acs%20csv_standardsmanual.ashx. value of morbidity and mortality (M&M) conferences: are Accessed March 4, 2019. minor complications important? J Surg Educ. 2012;69:326-329. 32. Charles R, Hood B, Derosier JM, et al. How to perform a root 18. Phillips DR, Ellsperman SE, Matt BH, Zarzaur BL. Comparison cause analysis for workup and future prevention of medical of automated post-tonsillectomy bleed capture with self-report. errors: a review. Patient Saf Surg. 2016;10:20. JAMA Otolaryngol Head Neck Surg. 2017;143:764-768. 33. Kruskal JB, Siewert B, Anderson SW, Eisenberg RL, Sosna J. 19. Huang Y, Fried LF, Kyriakides TC, et al. Automated safety Managing an acute adverse event in a radiology department. event monitoring using electronic medical records in a clinical Radiographics. 2008;28:1237-1250. trial setting: validation study using the VA NEPHRON-D trial. 34. Vollmer CM Jr, Sanchez N, Gondek S, et al. A root-cause analy- Clin Trials. 2019;16:81-89. sis of mortality following major pancreatectomy. J Gastrointest 20. Taggart M, Chapman WW, Steinberg BA, et al. Comparison Surg. 2012;16:89-102. of 2 natural language processing methods for identification of 35. Charles R, Hood B, DeRosier JM, et al. Root cause analysis bleeding among critically ill patients. JAMA Netw Open. 2018; and actions for the prevention of medical errors: quality 1:e183451. improvement and resident education. Orthopedics. 2017;40: 21. Rochefort CM, Buckeridge DL, Tanguay A, et al. Accuracy e628-e635. and generalizability of using automated methods for identify- 36. Univeristy of Michigan Center for Healthcare Engineering and ing adverse events from electronic health record data: a valida- Patient Safety. Triggering questions for root cause analysis. tion study protocol. BMC Health Serv Res. 2017;16:147. https://cheps.engin.umich.edu/wp-content/uploads/sites/118/ 22. Peterfreund RA, Driscoll WD, Walsh JL, et al. Evaluation of a 2016/08/RCA-triggering-questions-April-2015.pdf. Accessed mandatory quality assurance data capture in anesthesia: a March 4, 2019. secure electronic system to capture quality assurance informa- 37. VA National Center for Patient Safety. Glossary of patient tion linked to an automated anesthesia record. Anesth Analg. safety terms. https://www.patientsafety.va.gov/professionals/ 2011;112:1218-1225. publications/glossary.asp. Accessed March 4, 2019. 23. Driscoll WD, Columbia MA, Peterfreund RA. Awareness 38. US Food and Drug Administration. MAUDE—Manufacturer during general anesthesia: analysis of contributing causes and User Facility Device Experience. https://www.patientsafety aided by automatic data capture. J Neurosurg Anesthesiol..va.gov/professionals/publications/glossary.asp. Accessed March 2007;19:268-272. 4, 2019. 24. Bruce J, Russell EM, Mollison J, Krukowski ZH. The mea- 39. Gano DL. Apollo Root Cause Analysis: A New Way of surement and monitoring of surgical adverse events. Health Thinking. Moorseville, NC: Apollonian Publications; 1999. Technol Assess. 2001;5:1-194. 40. Marx DA, Watson J. Maintenance Error Causation: Phase IX 25. Walsh TA, Kydonaki K, Lee RJ, et al. Development of process Progress Report on Human Factors in Aviation Maintenance and control methodology for tracking the quality and safety of Inspection. Washington, DC: Federal Aviation Administration; pain, agitation, and sedation management in critical care units. 1999. Crit Care Med. 2016;44:564-574. 41. Abujudeh HH, Bruno MA, eds. Quality and Safety in 26. Jones S, Mullally M, Ingleby S, Buist M, Bailey M, Eddleston Radiology. Oxford, UK: Oxford University Press; 2012. JM. Bedside electronic capture of clinical observations and 42. Karuppusami G, Gandhinathan R. Pareto analysis of critical automated clinical alerts to improve compliance with an Early success factors of total quality management: a literature Warning Score protocol. Crit Care Resusc. 2011;13:83-88. review and analysis. TQM Magazine. 2008;18:372-385. 27. Lemon V, Stockwell DC. Automated detection of adverse 43. Johna S, Tang T, Saidy M. Patient safety in surgical residency: events in children. Pediatric Clin North Am. 2012;59:1269- root cause analysis and the surgical morbidity and mortality 1278. conference—case series from clinical practice. Perm J. 2012; 28. Saikali M, Tanios A, Saab A. Evaluation of a broad-spectrum 16:67-69. partially automated adverse event surveillance system: a poten- 44. Ramanathan R, Duane TM, Kaplan BJ, Farquhar D, Kasirajan tial tool for patient safety improvement in hospitals with lim- V, Ferrada P. Using a root cause analysis curriculum for ited resources [published online November 21, 2017]. practice-based learning and improvement in general surgery J Patient Saf. doi:10.1097/PTS.0000000000000442 residency. J Surg Educ. 2015;72:e286-e293. 29. Houliston BR, Parry DT, Merry AF. TADAA: towards auto- 45. Ferraro K, Zernzach R, Maturo S, Nagy C, Barrett R. Chief of mated detection of anaesthetic safety. Methods Inf Med. 2011; residents for quality improvement and patient safety: a recipe 50:464-471. for a new role in graduate medical education. Mil Med. 2017; 30. Zenati MA, Leissner KB, Zorca S, Kennedy-Metz L, Yule SJ, 182:e1747-e1751. Dias RD. First reported use of team cognitive workload for 46. Wong BM, Goldman J, Goguen JM, et al. Faculty-resident root cause analysis in cardiac surgery [published online ‘‘co-learning’’: a longitudinal exploration of an innovative December 19, 2018]. Semin Thorac Cardiovasc Surg. doi:10. model for faculty development in quality improvement. Acad 1053/j.semtcvs.2018.12.003 Med. 2017;92:1151-1159. 31. American College of Surgeons. Optimal resources for chil- 47. Littlefield M, Shah M. Closed loop quality management: inte- dren’s surgical care v1. https://www.facs.org/~/media/files/ grating PLM and quality management. https://www.plm.auto  10976817, 2019, 6, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599819878683 by Scholarly Information Univ Lib, Wiley Online Library on [31/08/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Balakrishnan et al 921 mation.siemens.com/de_ch/Images/AI-SiemensClosedLoop Agency for Healthcare Research and Quality; 2015. AHRQ Quality-ML-07_tcm782-117511.pdf. Accessed March 4, 2019. publication 15-0060-EF 48. ASQ. The define measure analyze improve control (DMAIC) 54. Kellogg KM, Hettinger Z, Shah M, et al. Our current approach process. https://asq.org/quality-resources/dmaic. Accessed March to root cause analysis: is it contributing to our failure to 4, 2019. improve patient safety? BMJ Qual Saf. 2017;26:381-387. 49. Scott SD. The second victim phenomenon: a harsh reality of 55. Peerally MF, Carr S, Waring J, Dixon-Woods M. The problem health care professions. Agency for Healthcare Research and with root cause analysis. BMJ Qual Saf. 2017;26:417-422. Quality Patient Safety Network website. https://psnet.ahrq.gov/ 56. Institute for Healthcare Improvement. Failure modes and effects perspectives/perspective/102/the-second-victim-phenomenon- analysis (FMEA) tool. http://www.ihi.org/resources/Pages/Tools/ a-harsh-reality-of-health-care-professions. Accessed March 4, FailureModesandEffectsAnalysisTool.aspx. Accessed March 4, 2019. 2019. 50. Nussenbaum B, Chole RA. Rethinking morbidity and mortality 57. Siemens PLM Software. How to conduct a failure modes and conference. Otolaryngol Clin North Am. 2019;52:47-53. effects analysis (FMEA). https://polarion.plm.automation.sie 51. Morrison RJ, Bowe SN, Brenner MJ. Teaching quality mens.com/hubfs/Docs/Guides_and_Manuals/Siemens-PLM-Pol improvement and patient safety in residency education: strate- arion-How-to-conduct-a-failure-modes-and-effects-analysis-FM gies for meaningful resident quality and safety initiatives. EA-wp-60071-A3.pdf. Accessed March 4, 2019. JAMA Otolaryngol Head Neck Surg. 2017;143:1069-1070. 58. Bagian JP, Gosbee JW, Lee CZ, Williams L, McKnight SD, 52. DeRosier J, Gosbee JW, Bagian J. Code UMHS-2014X: Root Mannos DM. The Veterans Affairs root cause analysis system Cause Analysis. Ann Arbor, MI: University of Michigan; in action. Jt Comm J Qual Improv. 2002;28:531-545. 2015. MLearning Modules (internal use only) 59. Institute for Healthcare Improvement. RCA2: improving root 53. Knox L, Levine J, Sommers B, Michaels L, Taylor EF. cause analyses and actions to prevent harm tool. http://www Primary Care Practice Facilitation Curriculum, Module 11:.ihi.org/resources/Pages/Tools/RCA2-Improving-Root-Cause- Using Root Cause Analysis to Help Practices Understand and Analyses-and-Actions-to-Prevent-Harm.aspx. Accessed March Improve Their Performance and Outcomes. Rockville, MD: 4, 2019.