Emergency department crowding in the Netherlands; evaluation of a real-time ambulance diversion dashboard

Abstract

Background

Emergency department (ED) crowding is a growing concern worldwide and associated with negative effects. In 2013, 68% of Dutch ED-managers experienced crowding on several days of the week. This resulted into the introduction in phases of an ambulance diversion dashboard, in order to influence ED input. Increasing numbers of Dutch EDs have implemented this dashboard, visualizing regional ambulance diversions by means of a traffic light.

Methods

This is a descriptive study of a nationwide online survey of Dutch EDs, conducted between January and October 2023. It included both qualitative and quantitative questions. The outcomes and analysis are derived from descriptive data of respondents’ experience of crowding as well as their usage and perceived effectiveness of the ambulance diversions dashboard.

Results

At the time of the survey, 62 of 82 Dutch EDs (75.6%) actually used the dashboard, of which 56 EDs responded (90.3% response rate). 69.7% Of ED managers experienced ED crowding more than three times a week. Of the respondents using the dashboard, 52.8% reported it only occasionally alleviates ED inflow. The purported reasons are the limited number of patients affected by the red light (ambulance diversion) and the presence of regional crowding. The effects of the orange light (impending ambulance diversion) on ED input differ greatly among hospitals, mostly due to their own internal agreements. In accordance, many respondents (53.6%) expressed dissatisfaction with the resources available to them to alleviate crowding.

Conclusion

After conducting a national survey, ED crowding is reported as a persisting nationwide problem with its prevalence largely unchanged since the introduction of the ambulance diversion dashboard. Most hospitals reported having insufficient resources to alleviate it. The effects of the ambulance diversion dashboard to decrease crowding are apparently limited because it affects a small portion of total ED presentations and because of the influence of regional crowding. The main function of the orange light is to increase ED throughput and output rather than reducing ED input.

Introduction

Emergency Department (ED) crowding is defined as a situation where the demand for emergency services exceeds capacity, leading to delayed care [1, 2]. This delay is associated with a number of negative effects like an increased workload, adverse events, mortality, reduced ED efficiency and lower patient satisfaction [3, 4]. Due to its negative effects, crowding is a growing concern globally [2, 5]. Dutch EDs are no exception to this concern, even though they are viewed as relatively efficient, compared to other European countries [6,7,8,9]. 68% of Dutch ED-managers reported crowding on multiple days in 2013, a trend expected to rise due to anticipated reorganizations in Dutch healthcare [7, 9].

A commonly used ED crowding model by Asplin et al. divides ED crowding into three interdependent components: input, throughput and output [10]. Many strategies have been implemented to address crowding by influencing these components, including ambulance diversions. Ambulance diversions, aiming to decrease ED crowding by managing patient input, was first described in the 1990s in the U.S. and has been widely used ever since [11, 12]. However, research on its effectiveness remains mixed, with concerns about its impact on transport times, care delays and increased mortality in certain populations [12,13,14,15,16].

In the Netherlands, ambulance diversions are managed by using a real-time online dashboard, developed by the Landelijk Platform Zorgcoördinatie (LPZ). This dashboard uses a traffic light system to indicate ED capacity: green for no capacity issues, orange for an impending closure, red for ambulance diversion and black for closure due to calamities such as technical failure or fire outbreak. Authorized users can adjust the system to communicate (impending) diversions, including diversions for specific patient categories like cardiac or obstetric care). In the dashboard, they are obliged to choose one out of five predetermined reasons for the ED closure (see appendix 1) and the expected duration. Ambulance diversions have a standard duration of one or two hours, depending on regional agreements. Situation-dependent, the diversions can be extended or shortened. However, proclaiming an ambulance diversion remains a team decision, without universal criteria on a national level. The aim of this dashboard is to effectively employ the available regional capacity of the individual EDs and to improve communication between regional partners.

The Netherlands is geographically organized in 10 so called ROAZ-regions (Regionaal Overleg Acute Zorgketen), each are collectively responsible for the organization of acute care in its region [17]. Hospitals from the same ROAZ-region and the Ambulance Service Control Room (ASCR) can consult the dashboard and thereby have insight into the capacity of nearby EDs. Furthermore, the dashboard displays a real-time crowding score based on the (m)NEDOCS ((modified)National Emergency Department OverCrowding Scale) that uses automatically collected data from the digital hospital system [18]. The aim of the dashboard is to optimally utilize regional ED capacity and thereby reduce crowding, by displaying and sharing this information. In this article we will evaluate the usage of the dashboard by a nationwide survey conducted among its users. Also, to contribute to the body of evidence of ED crowding and the use of ambulance diversions, the experienced effectiveness of the ambulance diversions and levels of perceived crowding are queried in this survey.

Methods

This descriptive study is based on a nationwide survey. The survey was developed in consultation with experts, including ED physicians, nurses, and managers. The final digital survey was distributed to all Dutch EDs and their managers through designated contacts in each ROAZ region. The survey, conducted in Dutch, included questions about ED characteristics, weekly experienced ED crowding, and local use of the ambulance diversion dashboard. Topics covered included internal protocols for announcing diversions, perceived consequences of ambulance diversion, and measures taken to improve patient flow and reduce the need for diversions. The full survey can be requested via email. To ensure the survey accurately reflected practice, it was completed by the primary users of the dashboard, with responses based on consensus. If multiple responses were submitted from a single hospital, the data were combined and consensus was reached by the authors, who verified it with the respondents. Non-respondents and EDs not yet using the dashboard at the time of data collection were asked to complete an abridged version of the survey, which included only two questions: annual patient volume and perceived crowding. Data collection took place between January and October 2023, followed by analysis. Qualitative responses were independently categorized by two researchers, and consensus was reached through discussion.

Results

All 82 Dutch EDs were contacted and received reminders between January and October 2023. A total of 62 EDs, actively using the dashboard at the initiation of the survey, received the complete survey. 56 Of them fulfilled the complete survey (90.3% response rate). Six non-responders from this group and 20 EDs, not (yet) actively using the dashboard, received the abridged version (Fig. 1). 20 Of the 26 contacted EDs fulfilled the abridged version.

Responder characteristics

Responder characteristics are shown in Table 1. The survey was predominantly completed by ED-managers (67.9%), followed by emergency physicians (21.4%), coordinating ED nurses (3.6%) or by a combination of these roles (7.1%). Two of the 56 (3.6%) responding EDs treated over 40.000 patients annually, while 50 (89.3%) were located in a general hospital and 52 EDs (92.9%) had no limited opening times. Supplemental Table 1 shows the response rate per region.

Dashboard usage

Dashboard usage defined by the authors includes actively using the red light in order to divert ambulances. Three of the 56 respondents did not meet this criterion. In two cases because of their rural location and one ED only treated self-referrals or patients referred by GPs. Therefore, these hospitals were excluded from further analysis, the following data is based on results from 53 hospitals. Of the remaining 29 EDs, 6 were non-responders.

Many differences were reported in the way the dashboard is utilized (Fig. 2). Thirty-one hospitals (58.5%) use the option of an impending ambulance diversion (orange light). Thirty-seven hospitals (69.8%) utilize partial stops for specific patient categories. These partial stops are notified by means of the dashboard but are often additionally communicated by phone to the on-call medical specialists and the ASCR. The predetermined reasons to announce an ambulance diversion are shown in appendix 1. Forty-seven hospitals (88.7%) have made additional internal agreements regarding the announcement of ambulance diversions. Twelve of the 53 respondents (22.6%) have stated they are using even other reasons besides the predetermined by LPZ. Most frequently given reasons being staff shortages (n = 6) followed by the lack of available inpatient beds (n = 4).

In 14 hospitals (26.4%) the decision to divert ambulances is made by one person only (emergency physician or coordinating ED-nurse) and in 39 hospitals (73.6%) this is a team decision. In most cases the team consisted of a coordinating doctor (predominantly an emergency physician) and the coordinating ED-nurse (n = 34, 64.1%). The predetermined duration of an ambulance diversion varied between 1 and 2 h (respectively in 28.3% and 71.7%) and is often regionally determined. Thirty-five of the respondents (66.0%) stated they “always” (50.9%) or “very often” (15.1%) actively discontinued the ambulance diversions if ED crowding is perceived to have subsided, instead of letting the timeframe pass by.

Effect of the dashboard – red light

Twenty-five respondents (47.2%) stated that the dashboard beneficially influences the inflow of patients. No respondents stated that the traffic light doesn’t influence the inflow at all. The remaining 28 (52.8%) stated that the traffic light is only occasionally successful in reducing ED inflow. The main reasons stated were the limited number of patients to which diversions apply and the fact that oftentimes the entire region is crowded, resulting in only a shift of the problem to other EDs (Table 2). In three regions, the majority of respondents (66.7%) stated that the traffic light only occasionally alleviates crowding (supplemental Table 2) These regions are among the most densely populated areas in the Netherlands with a large number of hospitals situated in the area.

Effect of the dashboard – orange light

Among the 31 hospitals using the orange light, the effects differ greatly (Table 3). In many hospitals, an internal protocol is activated in case of a proclaimed orange light. Providing EDs with a tool to predominantly influence ED through- and output instead of input, for example, by scaling up the ED staff, the initiating fast track and expediting admissions.

Perceived crowding

Seventy-six respondents (93% of all EDs) completed a question regarding ED crowding. Fifty-three of the respondents reported that ED crowding occurs three or more times a week (69.7%). In those EDs experiencing crowding less than three times a week, most treat a relatively small number of patients annually (69.6% treat < 20.000 patients, Table 4). A majority of the respondents (57.6%) lack resources to regulate or alleviate crowding. Reasons given are shown in supplemental Table 3.

Discussion

Frequent ED crowding seems to be a nationwide problem in the Netherlands as 69.7% of the respondents, experience ED crowding three or more times a week and over half of the respondents (54.5%) stated to have inadequate resources to alleviate it. This percentage on experienced crowding is in accordance with the calculated percentage of 68% in 2013 despite certain reorganizations in Dutch acute healthcare and the introduction of the ambulance diversion dashboard [7]. Hypothetically, increased attention to ED crowding of hospitals as well as the government and subsequent measures might have prevented a further increase of ED crowding over those ten years. To our knowledge, this is the first survey evaluating the Dutch ambulance diversion dashboard amongst its users. With a response rate of 90.3%, the results of this survey are believed to be an accurate representation of common insights and practice in the Netherlands. At the time the survey was conducted, 64.6% of all 82 Dutch EDs actually used the dashboard. The majority of respondents (52.8%) stated the dashboard only occasionally reduces patient inflow. Key reasons cited include the limited number of patients affected by the diversion and concurrent regional ED crowding. Ambulance diversions often shift the problem to nearby hospitals, potentially exacerbating regional overcrowding [15, 19]. Not questioned in the survey but known to the authors are different strategies applied to deal with this. In some regions, surrounding EDs need each other’s permission to announce an ambulance diversion, other regions have maximized the number of simultaneously announced diversions. Some respondents indicated that the ambulance diversion dashboard only occasionally reduces ED inflow due to misinformation and non-compliance with diversion protocols. A notable finding is the variability in how the impending diversion (orange traffic light) is interpreted, both across regions and within the same region. For some EDs, the orange light allows for diversion of specific patient categories, while for other EDs, its primary function is activating internal protocols to mobilize resources to alleviate crowding. Thus, the orange light is more commonly used to influence ED throughput rather than input. Several respondents noted that the orange light has minimal impact on ED crowding and primarily serves as a signal to staff and ambulances.

International literature on online capacity dashboards and queuing models to optimize regional ED capacity suggests that they can successfully reduce ambulance diversions [20,21,22]. However, comparisons across healthcare systems are challenging due to substantial differences. Unlike the Dutch system that allows to divert based on more subjective, human factors, many international dashboards use preset triggers, based on hard outcome measures. Furthermore, they often employ active dispatching strategies to divert ambulances to less crowded hospitals, since randomly diverting ambulances provides minimum benefit [22]. One effective queuing method allowed diversion of all patients, regardless of urgency - a strategy that is likely impractical and unsafe in practice [22].

In most surveyed hospitals, the decision to announce an ambulance diversion is made by a team of medical professionals, based on predetermined criteria set by the dashboard. However, 22.6% of the respondents noted that in reality these criteria do not fully capture the reasons for diversion. Since the diversion process remains subjective and influenced by human factors, as there are no universal rules or regulations. Therefore, the ambulance diversion dashboard appears to reflect perceived crowding rather than purely objective measures. This suggests that ED crowding is a complex issue, shaped by both subjective and objective factors, a view also supported by international studies showing significant variation in ED workload and diversion practices [23]. Another aspect of this complexity is the connection between EDs and the entire emergency medicine system. Research suggests that reductions in ambulance diversion are associated with increased ambulance offload delays if other factors contributing to ED crowding are not addressed [24,25,26]. Focusing solely on diversion status, without improving overall hospital throughput, may exacerbate ED crowding and negatively impact the broader emergency medicine system. This supports the notion that ED crowding is a multifactorial issue, requiring a multidisciplinary, systems-based approach [25, 26]. Merely using ambulance diversions to decrease ED crowding or to benchmark EDs is not comprehensive enough. Therefore, the authors recommend further investigations into negative consequences of (perceived) ED crowding, such as ambulance queuing and offload delays and their potential correlations with the Dutch traffic light system. Future ambulance diversion dashboards should reflect this complexity.

Limitations

With a 90.3% response rate among the presumed 62 EDs using the dashboard, the survey likely reflects common practices and insights. However, more EDs have adopted the dashboard since the survey, and the findings may not fully capture its usage among new users. The study’s reliance on opinion-based and partially qualitative data introduces potential reporting biases and incomplete responses. While participants were instructed to provide consensus-based input, most responses were submitted by a single individual, and six hospitals contributed multiple entries, requiring the authors to consolidate and verify a consensus, potentially influencing results. Additionally, internal reasons for diversion were not detailed, and the study exclusively focused on the ED perspective, omitting input from ambulance services and the ASCR.

Conclusion

A significant majority (69.7%) of Dutch ED managers report crowding at least three times weekly, with over half expressing dissatisfaction with available resources to address it. Most Dutch EDs have adopted an online ambulance diversion dashboard to optimize regional emergency care capacity. However, its impact on ED crowding is limited due to the restricted patient categories to which the diversion applies and the multifactorial nature of crowding, which extends beyond ED-input. Furthermore, ED crowding is often a regionally problem and the main focus of ambulance diversions is a local rather than a regional one. However, the potential negative effects of ambulance diversions might not be limited to an individual ED but influence the emergency medicine system as a whole. In practice, the dashboard might be more a reflection of perceived crowding rather than an objective crowding tool. The value of the dashboard may be more to influence ED throughput and output through initiation of increased hospital resources. This all supports the notion that ED crowding is a multifactorial and complex concept. Future research should investigate the relationship between objective crowding measures, perceived crowding, ambulance diversions and the effects on the entire emergency medicine system, to develop more effective, system-wide solutions.

Fig. 1

Responder flowchart. Detailed legend: Flowchart showing the collection of data among hospitals that use or do not (yet) use the dashboard

Fig. 2

Hospitals that use the dashboard. Detailed legend: Showing the distribution of the dashboard usage

Table 1 Responder characteristics

Table 2 Reasons why ambulance diversions (red light) only occasionally alleviate ED crowding according to respondents (n = 28)

Table 3 Consequences of an impending ambulance diversion (orange light) according to respondents (n = 31)

Table 4 Reported ED crowding by annual ED volume and type of hospital

Appendix 1

Predetermined reasons for announcing an ambulance diversion.

• Peak in patient presentations – a peak in the total number of patients presenting to the ED demands too much from the personnel capacity that optimal care can no longer be provided to new patients.

• Peak in high-acuity patients – a peak in the number of high-acuity patients that demand too much from the personnel capacity that optimal care can no longer be provided to new patients.

• Insufficient treatment capacity – Required quality of care cannot be provided because of insufficient treatment capacity (treatment rooms/equipment).

• Output blockage – Discharge from the ED to hospital wards of regional partners is impaired resulting in increased ED crowding.

• Internal calamity/technical issues – an internal calamity or technical problem which prevents the ED from providing adequate care.