Introduction
Following disasters, there is generally interest in identifying the processes and factors that would improve outcomes in future events. These “lessons learned,” or more precisely opportunities for improvement, are valuable insights into developing plans and responses that could provide future benefit.
Organizations have used after-action reviews (AARs) to try to identify lessons learned for years.Reference Savoia, Agboola and Biddinger 1 The US Army developed a formal process for conducting AARs in the 1990s.Reference Baird, Henderson and Watts 2 The use of AARs after disasters is commonly used in emergency management. As hospital emergency management has evolved over the past decades, hospitals have adopted many of the management processes used in non-medical emergency management fields. Regulatory requirements have also progressed and hospitals are now mandated to utilize AARs for both disaster exercises and real events. 3
The development of the US Homeland Security Exercise and Evaluation Program (HSEEP) helped formalize and improve this process across different types of agencies in the United States. 4 As part of the HSEEP formal AAR, an improvement plan, often called an improvement matrix, is devised with identified areas targeted for improvement. In this matrix, the problem and the individuals directly responsible for its resolution are identified with a timeline for resolution. The HSEEP program identifies a more structured approach to implementing recommendations of the AAR but doesn’t overcome the institutional barriers to improvements.
A recurrent theme in disaster reviews is that processes identified as needing improvement still fail in the same way in subsequent events. This was systematically evaluated in the article titled “Lessons We Don’t Learn: A Study of the Lessons of Disasters, Why We Repeat Them, and How We Can Learn from Them.”Reference Donahue and Tuohy 5 This study details the problems with making significant, sustainable improvements after disasters. One problem identified is maintaining the focus of leadership to make changes in the face of competing demands over time as memory of the event fades. For developing response plans for complex events, many of the needed after-action improvements require multi-disciplinary coordination for improvements across multiple agencies and departments. The field of process improvement (PI) includes techniques that lend themselves to analyzing and developing improvements in the field of emergency management. For example, the use of a rapid improvement event (RIE) described in PI literature is a technique that can help with developing sustained improvements in complex environments. The use of a RIE (also known as a “Kaizen Blitz”) is a technique used to rapidly work across groups on a well-defined problem.Reference Martin, Greenhouse, Kowinsky, McElheny, Petras and Sharbaugh 6 The RIE brings together representatives from multiple different functions and levels in the organization to focus on a problem or to improve a process. This event typically involves: (1) planning and preparation; (2) defining the current state and identifying inefficiencies; (3) designing improvements to eliminate inefficiencies and improve process reliability; and (4) following up to ensure that improvements are sustained rather than temporary.Reference Martin, Greenhouse, Kowinsky, McElheny, Petras and Sharbaugh 6
Report of the Information Technology Event
Modern hospitals are very dependent on information technology (IT) systems to function. 7 Over the past decade, US hospitals have transitioned to electronic health records with integrated laboratory and radiology systems. As such, unplanned IT failure represents an internal disaster threatening patient care and other operations across the organization. This threat is increasing with the advent of profit-driven IT attacks, such as ransomware.Reference Sittig and Singh 8 Developing an organized, system-wide, integrated response to this type of disaster requires an innovative way of performing an AAR and implementing its recommendations.
The University of Colorado Hospital (Aurora, Colorado USA), an urban academic center with 611 beds and 100,000 annual emergency department (ED) visits, experienced a near complete IT loss for 10 hours, which significantly impacted patient care. This outage affected the entire hospital, including most telephones and all computers. Patient care, especially for services of high computer use such as laboratory, radiology, and registration, were predominantly affected. The outage also particularly impacted areas such as the ED, operating rooms, and procedural areas. Many planning assumptions about reverting to “paper” processes were challenged by the large number of changes needed immediately, coupled with the lack of staff familiarity. This impeded prompt patient care and was stressful for staff involved. The hospital’s incident management system was overloaded with the detailed tasks required for effective response, and it was clear that major modifications in downtime response were needed.
In review of the events of the IT outage, the use of a traditional AAR followed by an improvement plan was felt to be insufficient to develop the needed corrective processes in a timely and sustainable fashion. This outage crossed multiple hospital departments and functional processes and would take years to change if approached in a sequential fashion.
Typically, the AAR assigns future improvements to be made but doesn’t conduct real-time decision making effecting these changes. The preliminary ED review showed key decision makers needed to be involved in the large formal review to accelerate change. Therefore, it was decided to use the technique of a RIE to conduct the review and improvement process.
Process Improvement Process
A RIE was performed focusing on the hospital ED processes with results designed to be applicable throughout the hospital. The RIE consisted of a structured preparatory phase, a two-day participatory phase with key leadership present to make immediate decisions, and was followed by a dissemination and implementation phase.
Preliminary work of the current state of downtime response in the ED was determined by the departments most involved in the IT outage. Planning and process mapping meetings were held weekly by key participants for five weeks prior to the scheduled RIE. The ED, laboratory, radiology, respiratory therapy, inpatient admission teams, and consultants documented their current downtime ED workflows and developed presentations with discrete nodes representing the critical processes. In addition, each of these departments was responsible for identifying stakeholders with the ability to make decisions and commit resources to support the desired outcomes from the RIE. An example of the previous state of the somewhat vague downtime radiology ordering is in Table 1.
Downtime Radiology Example Prior to Improvement Event
Abbreviations: ED, emergency department; MD, medical doctor.
During the RIE, an experienced PI facilitator methodically guided the work group through the processes involved in hospital ED downtime using the previously developed maps of current state. Major difficulties and points of failure were noted and organized into various “themes.” Each major area of concern (eg, obtaining imaging studies) was reviewed from start to finish of the process (eg, x-ray desired to result on chart). A future state was then defined and agreed upon by all stakeholders.
Very detailed hospital plans were developed for processes of downtime registration, patient flow, laboratory testing, and radiology. Additionally, the process for obtaining specialty consults and admitting patients to the hospital during IT downtime were developed. For each process, a detailed written description was added to the Downtime Annex to the ED Emergency Operations Plan and flowcharts, or pathways were developed with pictorial descriptions of the response plans (Figure 1). These templates are now in use in the ED and are undergoing revision for internal hospital use for future unplanned IT downtimes.
Emergency Department Radiology Workflow. Abbreviation: ED, emergency department.
The RIE outcomes had pre-defined sustainability plans. For example, the entire downtime process including whiteboards, bed numbered chart clipboards, paperwork, and instructions are stored in easily accessible areas in the ED. Further, the RIE has been followed by structured staff drills on downtime methodology. This drill process uses “microdrills” where a trainer has staff briefly review the processes and the developed templates. Staff also review where the downtime materials are stored for emergencies. Additionally, functional exercises (complete setups of the downtime process and supplies) implementing the downtime procedures were designed and are used to validate the overall downtime process. The global hospital training and exercise plan has been modified to include downtime exercises in a structured fashion occurring at least quarterly, with monthly drills for each department. An accountability plan has been created to drive the continuation of these process practice drills.
Discussion
The use of PI methodology for AARs in the field of emergency management holds promise in achieving sustained improvements. The situations most likely to benefit from the resources allocated to a formal RIE are complex events that span multiple different functional groups or agencies. For these larger complex problems, PI techniques leverage leadership engagement to develop and implement improvements both rapidly and sustainably.
Limitations
The use of the RIE, while effective in achieving change, has some downsides. The time commitment of senior staff to participate in the planning and actual event is significant. This can be a major limitation in using PI methodology.
Conclusion
In the appropriate settings, RIE methodology can be used to effectively conduct emergency management AARs. This holds the potential for immediate and sustained improvements.
