Background
In The Netherlands, major changes in the organization of emergency care are planned to decrease health care costs. For example, the closure of 40% of the emergency departments (EDs) was recently discussed [
1]. This could impact care in EDs by causing crowding. Dutch health policy makers and insurance companies plan to integrate general practitioner (GP) cooperatives and EDs into one facility to prevent patients from self-referring to the ED. Both changes may affect ED patients’ length of stay (LOS) and crowding.
Although ED crowding is not yet a major problem in this country according to expert opinion [
2], anecdotal evidence suggests that current ED patients experience a longer LOS compared to some years ago, which is indicative of ED crowding. However, no multicenter studies have been performed to quantify LOS and assess crowding at Dutch EDs. We conducted this study to describe the current state of EDs in The Netherlands, including ED characteristics, patients’ LOS, and ED nurse managers’ experiences of crowding. To study the effect of the planned changes in the organization of emergency care on ED patients’ LOS and ED crowding, we plan to repeat this study in 3 years.
Methods
Study setting and study design
In The Netherlands, there are 132 hospital locations. Ninety-nine hospital locations have EDs [
3], including 91 general hospitals and 8 university hospitals. There are an estimated 2.2 million ED visits annually [
4]. Basic health insurance is available to all citizens: half of health care is paid by taxes and employers, half by insurance. Most people are registered with a local GP. The presence of emergency physicians (EPs) is increasing [
5]. To date, there are almost 300 trained and registered EPs working in 80% of the EDs [
6].
A survey study in The Netherlands was performed in November 2012. The survey was addressed to the ED management; it could be completed by a nurse manager, staff nurse, medical manager, or EP. Surveys were returned to the primary investigator. Data were entered into SPSS 20.0 (SPSS Inc., Chicago, IL, USA). Institutional review board exemption was granted.
Study protocol
At the onset of this study, a letter announcing the survey was published on the website of the Dutch Association of ED Nurses (NVSHV) and was also noted by the national press. Surveys were distributed to all ED nurse managers using an address list published by the Ministry of Health [
7] combined with an address list obtained from the NVSHV. Included in the e-mail were: a letter explaining the survey, its purpose, and a digital version of the survey. A paper-based version of the survey and a second e-mailing with an online version of the survey were sent to non-respondent EDs in January 2013 to increase the response rate.
Survey content and definitions
A draft survey was created, and after consultation of experts (two EPs and two ED nurse managers), a final version was provided (Appendix).
The survey included questions regarding type of facility, hospital size, annual ED census (based on year 2011), change in volume of annual ED visits from 2008 to 2012, ED bed capacity, number of ED nurses and physicians per shift, patients’ LOS, percentage of self-referred patients (self-referrals), percentage of patients arriving by ambulance, admission rates, and how often ambulance diversion was used. Additional questions included how often crowding occurred, which time periods had the worst episodes of crowding, putative causes of crowding, and what measures had been undertaken to improve patient flow. Respondents chose from a list of causes of crowding and from a list of measures to manage crowding. Respondents were instructed to circle all appropriate answers, creating the possibility of more than one answer per respondent. Respondents were also provided the opportunity to fill in answers other than the answer lists provided. If actual data from hospital databases were not available, respondents were allowed to report estimations. They were also allowed to skip questions.
LOS was defined as the interval between patient registration and the moment the patient left the ED. Based on previous research, crowding was defined as having more patients in the ED than treatment rooms or more patients than staff should ideally care for [
8], and overcrowding was defined as dangerously crowded, with an extreme volume of patients in ED treatment areas forcing the ED to operate beyond its capacity [
9].
Data analysis
Data were reported as mean and standard deviation and median and ranges, in case of a skewed distribution. To investigate whether differences occurred by type of hospital, we examined the data for the overall group as well as for type of facility separately, using two-tailed t tests, Kruskal-Wallis test for ordered categories and Fisher’s exact tests where appropriate. Statistical significance was assumed at a level of p ≤ 0.05.
Discussion
LOS at EDs in The Netherlands (119 min for discharged patients, 146 min for admitted patients) is short compared to published LOSs in other countries [
10,
11]. In the USA, admitted patients may have an LOS of over 24 h during times of severe crowding [
12]. Despite this relatively short LOS, frequent crowding appears to be a Dutch problem according to our respondents, with 68% of them reporting that crowding occurred several times a week or even daily, and half of those reporting that, besides crowding, their ED was also overcrowded two or more times a week. Our findings are somewhat milder compared to studies performed in the USA more than 10 years ago by the co-authors [
13‐
15] in which 91% of the ED directors in the USA reported crowding to be a problem, probably indicating that crowding is better controlled in The Netherlands. However, if health restructures continue (closure of EDs and decreasing inpatient bed capacity), crowding may become more prevalent. Our respondents named several factors they believed to contribute to ED crowding, and their answers were similar to those from other international studies [
16‐
19]: consultation delays, shortages in ED space and beds, admission delays, shortages of acute care inpatient beds, lack of nursing staff, and laboratory and radiology delays.
In the Dutch lay press, it is suggested that the problem of crowded EDs is predominantly caused by inappropriate use of emergency services by patients seeking care for non-urgent problems. The same was suggested in the USA in the early 1990s in several position statements [
20]. Integration of GP with ED services has had mixed effects: unsuccessful in some hospitals in Australia and New Zealand, while effective in diverting patients in one study from The Netherlands [
21,
22]. This Dutch study did not measure effects on crowding. Current research on ED crowding suggests that discouraging the use of the ED for non-emergency issues will not solve the problem. Rather, the issue of output, for example, inadequate inpatient capacity for a patient population with an increasing complexity and severity of illness, is now believed to be the single most important factor contributing to ED crowding [
23]. Our respondents agreed: 64% cite hospital bed shortages as a problem contributing to crowding. Only 16% blamed self-referrals for crowding, while many (43%) believed crowding occurs when too many patients who are referred by the GP or multi-trauma patients present at the ED (33%). High patient acuity has been cited as a significant contributing factor to ED crowding [
13].
Besides GP cooperatives, numerous measures have been implemented to improve ED efficiency and alleviate crowding in Dutch EDs. These measures have been mentioned in the past international literature about ED crowding. Examples include implementing observation units [
24] and creating a fast-track unit [
25]. A few measures described in the international literature were rarely mentioned in our study, such as ambulance diversion. For many Dutch EDs, ambulance diversion is not an option, even when conditions warrant diversion. Most university and major EDs have no alternative treatment site, since EDs in The Netherlands have special assignments, such as a dedicated trauma center designation. For non-trauma ambulance patients, diversion would be possible; however, hospitals have strong economic pressures to remain open. Only one respondent reported requiring diversion several times per week.
The body of evidence documenting the adverse effects of crowding has grown up to the sky. Crowding not only compromises the quality of care, it also worsens clinical outcomes [
26] and has negative effects on staff satisfaction and health [
27]. It is apparent that most countries have been struggling with ED crowding for many years, and the focus has shifted from identifying causes and consequences to finding solutions. The Dutch are following this trend. Some Dutch EDs have implemented fasttrack (38%) with or without nurse practitioners (16% of the respondents use nurse practitioners), which has been reported to help decrease LOS [
28]. ED nurse managers recognize that the cause and solution to ED overcrowding lie outside the ED. They consider ED crowding as a system-wide problem instead of an ED phenomenon, as seen in other countries. Facilities are increasingly utilizing ED-managed overflow units (acute admission units, transit lounges and flexible beds) to make room for incoming patients. These overflow units mitigate crowding by giving the ED staff a way to control patient outflow to some extent [
29,
30]. Other important potential solutions, such as expediting discharge from the main wards, were not mentioned by our ED nurse managers.
Future studies in The Netherlands should focus on determining which aspects of restructuring healthcare are most closely related to ED crowding. The Dutch can learn from what is already known in other countries with severe crowding. Despite environmental, demographic and healthcare organization differences among countries, the causes and consequences of crowding appear to be universal, and certain strategies will alleviate crowding wherever they are implemented. From the existing evidence, it is clear that multidisciplinary system-wide support is necessary to solve ED crowding. Introducing quality benchmarks in The Netherlands would be useful. Moreover, EDs should start collecting a uniform set of process measures that provides real-time observation of the operation of the department like the crowding measures recently identified by Beniuk et al. [
31]. This would facilitate across-facility comparisons to identify the best practices that work in our healthcare system.
Limitations
First, our survey has not been validated yet. As in most surveys, our results are subject to reporting errors, non-response, and incomplete responses. In The Netherlands, several different patient information systems are used, and hospitals use different definitions for the data that are tracked. For example, referral source and transport were used interchangeably at different sites: in some EDs, all patients brought in by ambulance were documented as 'ambulancepatients’, while in other EDs patients who were referred by a GP but transported by an ambulance were not registered as such. In some Dutch EDs, visits are not tracked, so a few respondents presented estimations instead of actual data. Although this data collection is far from ideal, we believe the benefits of multicenter participation outweighed the weaknesses of variation in operational data. We do not know if the characteristics of non-responding EDs were similar or systematically different from those of responding EDs. However, our purpose was not to assess the population as a whole but rather to describe the current status of EDs, current LOS, and ED nurse managers’ experiences of crowding.
Another major limitation is that no standard definition of ED crowding exists [
16,
32]. Several factors associated with crowding were included in the survey, but no standard method was used for actually defining crowding. ED crowding assessment tools (e.g., EDWIN [
33], NEDOCS [
34]) are not yet used routinely in The Netherlands. Some metrics that define patient throughput, such as ambulance diversion hours [
35] or the number of patients leaving without being seen [
36], are used as surrogate markers of crowding in the absence of a widely accepted definition [
37]. Measuring crowding with hours on ambulance diversion or with the percentage of patients leaving without being seen will not give a true picture of ED conditions in The Netherlands, since both circumstances are rare. As with other studies [
33,
38], we used staff perceptions of crowding. Although subjective, ED nurse managers’ sense of how their EDs operate was the closest accurate measure of current crowding. After national implementation of crowding measures into the ED information system in The Netherlands, further studies assessing ED crowding will be necessary, using empirical data to quantify ED nurse managers’ experiences.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
CvdL had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: CvdL, RR, NvdL. Acquisition of the data: CvdL, RR. Analysis and interpretation of data: CvdL, RR, NvdL. Drafting of the manuscript: CvdL, RR, RD, RL, JR. Critical revision of the manuscript for important intellectual content: RD, RL, CL, JR. All authors read and approved the final manuscript.