It has been well established that many emergency departments (EDs) are operating at or above capacity in many countries, including the United States, Canada, and Australia.1-6 Evidence is mounting that these capacity levels negatively impact care quality.7-13 In response, hospitals around the world are experimenting with a variety of patient flow strategies (eg, use of bed coordinators, advanced protocols, bedside registration, and full-capacity protocols) to reduce ED crowding, often with promising results.14-19 However, little evidence or guidance currently exists on how to implement patient flow improvement strategies and specifically on the factors that facilitate or hinder implementation.
The science of implementation is evolving, and throughout the past several years, an increasing amount of attention is being given to the process of implementing hospital quality improvement activities.20 There is greater recognition that understanding why or how a program works is as important to potential adopters as whether it works.21 However, one of the challenges to conducting implementation research is the literature bias-implementation failures are rarely reported.22 Although recent work has explored the factors that facilitate or hinder the implementation of quality improvement projects,23 evidence specific to the implementation of efforts to improve patient flow and reduce ED crowding is limited.
The purpose of this effort is to identify and describe the facilitators and barriers encountered by 6 hospitals in the United States that implemented strategies to improve patient flow and reduce ED crowding and to identify the successful tactics used to address the barriers. Findings from this small sample are not meant to be generalized to all hospitals. However, the identification of common facilitators and barriers experienced by these hospitals, as well as actions taken to successfully overcome barriers, may help to prepare other hospital and ED leaders for implementation of these or similar strategies at their own facilities. Our results may also be useful to the organizations that provide support to hospitals as they implement quality improvement efforts.
Participants and setting
In June 2008, we recruited 6 hospitals to participate in an 18-month learning network led by Urgent Matters (UM), a national program funded by the Robert Wood Johnson Foundation, dedicated to finding, developing, and disseminating strategies to improve patient flow and reduce ED crowding. These hospitals were selected from approximately 50 hospitals that responded to our recruitment letter, forwarded to them by state hospital associations. The 6 hospitals were selected to provide diversity in terms of size, ownership, teaching status, and safety net status. To be eligible, the hospitals had to state their willingness to accept the requirements of the learning network.
Specifically, since previous studies of implementation have shown the importance of collaborative teams, each hospital was required to form a patient flow improvement team consisting of a nurse, physician, additional representatives from the ED, senior hospital leader, research analyst, and other relevant team members, such as staff from ancillary services and information systems. Each team was required to select at least 1 strategy to improve patient flow and reduce ED crowding, develop an implementation plan, and submit monthly progress reports to the UM staff members, as well as performance data to support an independent, external evaluation of their efforts. Finally, each team agreed to participate in periodic web seminars, monthly teleconferences, 3 site visits by UM staff, and three in-person collaborative meetings.
Each hospital received a $10000 stipend to cover a portion of the costs associated with participation. The collaborative was launched in October 2008. The teams were offered examples of improvement strategies but were also told that they could select other strategies as long as they were designed to improve patient flow/reduce ED crowding, could be implemented within 3 months, and the impact could be measured. The strategies selected and implemented by the patient flow improvement teams are shown in Table 1. Across the 6 hospitals, 8 strategies were implemented. Initial implementation dates ranged from March 2009 to September 2009. Our assessment of the impact of the strategies was mixed. Although at least some respondents from all 6 hospitals speculated that their efforts resulted in improved patient flow, our analysis indicated a statistically significant reduction in ED length of stay or the percentage of ED patients that left without being seen at only 3 hospitals (hospitals B, C, and F) after the implementation of the strategies.
|Table 1 OVERVIEW OF STRATEGIES PROPOSED UNDER THE URGENT MATTERS LEARNING NETWORK|
We conducted 2 rounds of individual interviews with all members of the hospitals' patient flow teams and other staff identified by a team as serving a role in the implementation of a strategy. First-round interviews were conducted in person between August 2009 and September 2009. Three experienced qualitative researchers conducted the interviews in 2-person teams. A second round of interviews was conducted by telephone by 1 researcher during January 2010 and February 2010. We interviewed 75 individuals during the first round and 54 individuals during the second round, across all 6 sites.
During both rounds of interviews, respondents were asked several open-ended questions about the facilitators and challenges to the implementation of their strategies. We included prompts in the interview protocols to draw out more specific information about certain factors that may have facilitated or hindered implementation, including the availability of resources, use of formal improvement methods, staff and executive support for the strategies, and participation in the UM learning network.
The interview questionnaires were reviewed for clarity and content by an expert in qualitative research methods and a hospital leader who recently implemented a patient flow improvement strategy. Interviews were recorded and transcribed for accuracy. We used an inductive approach to analysis. Codes were derived after an initial review of the transcripts and weekly debriefings among the interviewers/researchers. Two researchers manually applied the codes to the transcripts independently. There was a high level of agreement, and any disagreements between the 2 coders were resolved by discussion until a consensus was reached. Themes, defined as concepts that frequently appeared, were identified from the analysis of the coded text. We shared our results with the leaders of the hospital patient flow improvement teams and incorporated their feedback.
The study design, interview questions, and informed consent procedures were approved by the institutional review board of the Health Research and Educational Trust.
Table 2 provides an overview of the facilitators and challenges to implementation reported by patient flow improvement teams from 2 or more of our 6 hospitals. In some cases, the teams developed successful approaches for addressing the challenges encountered (Table 3). In this section, we describe the common facilitators encountered during implementation and the challenges and approaches used to mitigate those challenges.
|Table 2 COMMON FACILITATORS (+) AND BARRIERS (-) TO STRATEGY IMPLEMENTATION, BY STRATEGY|
|Table 3 SUCCESSFUL APPROACHES USED BY URGENT MATTERS HOSPITALS TO MITIGATE BARRIERS|
Facilitators to the implementation of strategies
Participation in the UM learning network
Across all 6 hospitals, the most frequently cited facilitator to implementation was participation in the UM learning network. First, the network provided the hospitals with structure and a firm timeline for implementation. As a condition of participation, hospitals were required to provide UM staff with detailed implementation plans, which meant that the teams had to consider all of the intermediate steps needed to implement the strategy, the resources required, and the individuals who should be involved. Second, the external accountability created by UM helped to ensure that the proposed improvement strategies received appropriate attention despite other large projects underway during the same time (eg, large hospital construction projects and implementation of electronic health records).
Third, many respondents said that the prestige associated with being selected to participate in the UM learning network, as well as the $10 000 stipend, was beneficial to obtaining the attention and support of senior hospital leaders. Two project team leaders reported that participating in the learning network made the hospital leadership more receptive to ED requests for resources to support the improvement strategy. Finally, collaboration with other hospitals provided peer-to-peer networks that aided implementation. For example, representatives from another hospital offered hospital C networking suggestions to help to expand its nurse practitioner (NP) applicant pool.
Strategic selection of planning team members
Several strategies required collaboration with units or individuals outside of the ED, and many respondents discussed the benefit of an inclusive planning team. For example, at hospital F, the patient flow team approached inpatient managers and staff during the design phase for their input on a tool to expedite admissions from the ED. Although establishing inclusive teams may result in a longer planning process, collaborating with stakeholders from other areas of the hospital during the planning phase may be time saving over the long term. For example, at hospital A, the patient flow team did not confer with physicians when modifying the consult request process, and as a result, the team had to later redesign the process on the basis of physicians' feedback.
Also, having capable, adaptable, and willing staff pilot the changes proved useful to the launch of strategies at 2 hospitals. At hospital C, a small group of experienced nurses launched the open-bed strategy. Identifying supportive, willing nurses to test the strategy would permit the planners to make corrections to the strategy before it was rolled out to the rest of the nursing staff.
Executive support and the availability of resources
Executive support was present at most of the hospitals, often demonstrated by the addition of key resources needed to implement the strategies. For example, at hospital E, because of senior leaders' strong commitment to improving patient flow, the ED was able to hire an additional physician so that mid track, a special unit dedicated to expediting care for mid-acuity patients, could be staffed by a full-time physician.
Notably, at 2 organizations, requests for resources to support the implementation of the strategies were initially denied. However, several months after the implementation, it became clear that these strategies would not reach their full potential without the addition of resources. For example, at hospital B, staff members requested 2 new computers on wheels ($8000 each) to aid the new improved registration process. Although their request was initially denied by hospital leadership due to budgetary constraints, it was approved shortly after executives attended a UM collaborative meeting.
Staff-driven improvement strategies
Three of the strategies (standardized registration and triage at hospital B, 5-level triage at hospital C, and fast-track improvement at hospital D) were designed and led by staff. Respondents at all 3 organizations explained that implementing a staff-driven strategy minimized the amount of staff resistance that might have occurred from a top-down approach.
Staff engagement was facilitated by the use of Lean process improvement methods at hospitals B and C. Both hospitals formed multidisciplinary teams to map current processes, identify changes that would improve efficiency, test the changes over a short time period, and make adjustments until the goal was met. At hospital B, respondents reported that Lean tools resulted in a better understanding among nurses regarding why the changes were being made and a stronger strategy than would have been identified from a top-down approach.
An aligned reporting structure
Several respondents reported that an aligned reporting structure facilitated the implementation of their strategies. Most of the strategies involved coordination across multiple hospital units or staff with different roles. Having a common supervisor, who was supportive of the strategy, served to emphasize the importance of collaboration. For those strategies requiring coordination across hospital units, having a common senior leader who supported the strategy enforced compliance with the implementation plan.
At hospital D, the chief operating officer established a new position, vice president for emergency medicine clinical operations, responsible for overseeing all ED operations. This change was significant because for the first time, all ED physicians and nurses reported to the same individual. The new vice president was a proponent of patient flow improvement, and staff understood that improving the fast track was a priority for the department.
Implementation of simple process changes
Respondents from 2 hospitals said that they had little difficulty implementing their strategies because the changes did not represent a radical departure from previous processes. For instance, at hospital E, the implementation of mid track was a relatively straightforward change. The mid-track space was previously used as an observation unit, so physicians were already accustomed to directing patients to that area, and a NP continued to oversee operations in the space.
A flexible and robust information technology system
Two of the hospitals in the learning network had sophisticated health information technology (IT) systems and in-house expertise that facilitated implementation. For example, the patient flow improvement team at hospital A developed a new system for requesting specialty consultations. According to 1 respondent, the ED physicians "are thrilled with the new computerized process because it doesn't interfere with their workflow."
Barriers to the implementation of strategies
Staff resistance was the most common barrier encountered during implementation. Resistance often arose because the changes resulted in more work or a disruption in workflow or because frontline staff members were not included in planning. At hospital C, under the open-bed strategy, incoming patients were immediately directed to an open bed, when available, for triage and registration. Some of the staff nurses complained to the charge nurses about receiving more patients than they were accustomed to, and as a result, the charge nurses at times reverted back to the previous process.
The tool developed by the team at hospital F to facilitate communication and the transfer of patients between the ED and inpatient units was not well received by the progressive care unit (PCU), the cardiac step-down unit. The PCU staff members were concerned that the paper form did not contain all the information needed for patients with complex conditions.
There were a couple of approaches that the patient flow improvement teams at these hospitals and others used to successfully overcome staff resistance. For the open-bed strategy at hospital C, the ED director and assistant director made daily rounds to constantly reinforce the new policy. One respondent said that staff needed a clear message, "This is an expectation now. It is not a choice." The constant reinforcement was also paired with positive feedback, for example, "you guys did a great job yesterday, we had 115 patients, and only 6 of them left."
At hospital F, the implementation team made changes to address the concerns of PCU staff. They worked with the IT department to create an electronic version of the tool with more detailed information for complex patients.
Entrenched organizational culture
Culture has been defined as a set of shared values and beliefs that can influence the quality of care within health care teams.24 The culture within an organization, unit, or department can have profound effects on the receptivity of its members to change of any kind, particularly when a proposed change appears to run counter to the existing culture. This indeed occurred at some of the UM hospitals, and many of the patient flow improvement teams found it difficult to change attitudes or habits. Culture trumps strategy, and, as 1 respondent put it, "You have to change how people think."
For example, the fast track at hospital D had long been a source of frustration for ED leaders and staff. There was a general lack of respect for the fast track, and staff would monopolize fast-track equipment, supplies, and staffing when the ED was busy. One of the factors that helped to improve the fast track's reputation was use of the Lean process improvement tools. After a multidisciplinary team did a value stream map of the entire process, it became apparent that the many inefficiencies and long lengths of stay were because of higher-acuity patients being inappropriately assigned to fast track and a lack of staff and resources to handle the fast-track patient load rather than staff perceptions that NPs in the fast track were moving too slowly. Similarly, at hospital B, use of the Lean tools led nurses and registrars to a better understanding and appreciation of each other's work.
Still, changing culture was more challenging at some of the other organizations, and it was not easily overcome by any particular approach. Respondents from 3 different hospitals said that it was more challenging to change the thinking and habits of more experienced staff members who were fully ingrained into the organizational culture.
Lack of staffing resources
Several of the strategies required the addition of personnel, and some of the hospitals struggled with recruitment and hiring freezes. Because of the economic recession and overall financial pressures faced by participating hospitals, hiring additional staff was not an option for some organizations. At hospital D, it was clear to the team that a dedicated nurse and technician were needed to assist the NP in fast track. One respondent said, "We asked administration for an additional tech, but I can tell you they said 'no.'" An ED nurse and technician were pulled from the ED to staff fast track, though as 1 respondent put it, "We robbed Peter to pay Paul." Another respondent from hospital D said that one of the lessons learned was that there "needs to be dedicated personnel, and the strategy cannot be predicated on pulling people from the main ED."
Previous failures to improve patient flow
Respondents from 2 hospitals said that previous failures to implement or maintain quality improvements led to cynicism among some staff members about the strategies proposed under the UM learning network. In several cases, it was not the first time that many staff members had participated in various improvement strategies, and many staff members believed that the proposed changes were yet another exercise that would later be abandoned-management's flavor of the month.
Respondents identified 2 factors that alleviated the skepticism associated with the new interventions. First, constant reinforcement from department leaders demonstrated that the strategy would become standard operating practice. Second, site visits by UM staff and evaluators made it clear that outside organizations were monitoring progress and emphasized the importance of following through with the interventions.
Lack of data to monitor progress
The hospital teams needed information to track whether their changes were effective in improving patient flow or whether further refinements were needed. However, patient flow teams at 3 hospitals did not have access to data to assess their progress. Hospital B had a very sophisticated system to track patient flow; however, the ED nurse leader did not share the data with the implementation team. That barrier was never overcome. The information systems at hospitals D and A were not configured to capture data on the indicators needed to monitor progress. To address this shortcoming, both hospitals developed a manual process for data collection. While these data were considered to be "better than nothing," the data collection was time consuming and prone to error.
The 6 hospitals that participated in the UM learning collaborative encountered several common factors that facilitated or hindered implementation. Findings from this study, while not generalizable, may offer guidance to hospital and ED leaders who are about to implement patient flow improvement strategies. Before implementation, several steps should be taken to leverage the factors that facilitate implementation. Examples are asking staff to develop and implement the strategy, assembling an inclusive planning team, identifying simple process changes with the potential for a large impact, and identifying willing staff to pilot test the strategy. Other factors require a long-term effort to accomplish (eg, aligned reporting structure and robust IT system) but are nonetheless important for supporting many patient flow improvement projects. The number and diversity of factors needed for successful implementation indicate the need for a collaborative approach to process improvements. It is unlikely that ED administrative and clinical staff alone can obtain all of the facilitators identified in this study.
It is important to note that participation in the learning network was the most commonly cited facilitator to implementation. Working within the network compelled the participating hospitals to be accountable for results, making it difficult to abandon or change the strategic direction once it decided on a particular improvement strategy. Unfortunately, the learning network was open to only 6 hospitals. Yet, there may be other ways for individual hospitals to foster accountability and ensure that patient flow improvement projects receive the attention and priority that they deserve. For example, hospitals may tie ED flow metrics into executive compensation, include emergency flow metrics in a regular dashboard presented to the board, or post ED wait times on the hospital's Web site.
The learning network also fostered structure and adherence to timelines. These factors may be replicated by the use of formal quality improvement methods, such as Lean or Six Sigma. Furthermore, the collaborative learning opportunities that the learning network afforded might be obtained through participation in patient flow initiatives through professional societies (eg, American College of Emergency Physicians and Emergency Nurses Association) and other quality improvement organizations. Because these benefits were essential to successful implementations, organizations that support hospital quality improvement efforts should consider investing more resources in collaborative or virtual-collaborative models.
Hospitals and ED leaders should also be aware of the challenges that are likely to arise during implementation, particularly staff resistance and culture barriers. Our findings also illustrate that the presence of challenges does not necessarily mean that a strategy is doomed to fail. However, hospital leaders and improvement teams should be prepared to commit time and effort to supporting strategies even in the face of adversity. This study offers some examples of strategies to mitigate barriers, but further research is needed to identify effective approaches.
In sum, our results are consistent with findings from previous studies on the implementation of hospital patient safety or quality improvement projects. First, implementation of quality improvement efforts is challenging, and barriers are frequently encountered.25 Second, staff involvement, leadership support and commitment, and capable reporting and information systems facilitate implementation, while resistance to change, organizational culture, and lack of resources hinder it.23,26-27 Third, our results support the notion that learning networks can be effective in facilitating change.28
There are several limitations to this study that deserve mention. First, the study included only 6 hospitals. While the 6 participating hospitals are diverse, they are not nationally representative. In addition, these 6 hospitals self-selected into the collaborative and as a result might possess characteristics, including an openness to (or readiness for) change, that differentiate them from hospitals that either did not choose to participate or were not selected to participate. The improvement strategies included here do not represent a full menu of possible strategies. Also, the patient flow improvement teams' experiences were shaped by their participation in the learning network, which further limits the external validity of our findings. Hospitals adopting patient flow improvement strategies on their own in the absence of guidance from the UM technical advisors may face more or different challenges. Second, with little prior research to build upon, we selected a qualitative research design to identify common facilitators and barriers to the implementation of patient flow improvement strategies. The design prevents testing of hypotheses, such as whether certain facilitators or barriers have a stronger influence on implementation than others. This subject remains unexplored and would be appropriate for further research. Third, our article focuses on common facilitators and barriers (ie, those that were identified during the implementation of at least 2 different strategies). There may be additional strategy-specific challenges, such as state or federal regulatory barriers, that are uncommon but have a considerable impact on the implementation of strategies.
There have been many calls for hospital and ED leaders to adopt patient flow improvement strategies to address the widespread problem of crowded EDs. This study revealed several common facilitators and barriers that may be encountered during implementation. Focusing on the factors that facilitate implementation and developing a strategy to address common challenges may leave leaders better prepared to implement future patient flow improvement strategies.