The emergency department landscape in The Netherlands: an exploration of characteristics and hypothesized relationships

Nationwide optimization of the emergency department (ED) landscape is being discussed in many countries all over the world for reasons such as cost reduction and quality improvement. In The Netherlands, one such country, several guiding reports on this topic have been published [1–3]. The emphasis here was mostly on the number of EDs actually present at the time versus a proposed minimum number of EDs needed in the future at which sufficient ED accessibility to the population could be maintained [1–3]. The predominant idea was that by concentrating emergency care in fewer EDs costs would be saved and quality of care would increase. The 45-min standard (45-min time limit for ambulance services to deliver patients to an ED, including arriving at the patient within 15 min after being dispatched) applies as a restrictive measure for this sufficient accessibility from the Dutch government [4, 5]. In line with the development of last decades, the number of EDs has declined gradually from 105 in 2010 to 87 in 2016. In all cases, that was a result of mergers of hospital organizations [6]. However, whether the intended nationwide optimization goals of cost reduction and quality improvement are achieved is being questioned [7–10]. At the same time, pressure across the width of the acute care system has increased and in particular crowding in EDs with all its negative aspects has become a growing problem [11–13]. We assumed that reducing the number of EDs as strategy in itself has limited effectiveness. This might be among other things because in practice a broad variation of departments in many ways is hiding behind the commonly used term ED. The diversity within the ED landscape in The Netherlands has been demonstrated earlier [4, 6, 14]. However, structural insight into similarities as well as differences of ED characteristics is missing. This knowledge is needed to provide better steering information which will contribute to strategies aiming to optimize the ED landscape.

With this study, we aimed to provide additional insight in the ED landscape of The Netherlands in two ways. First is by providing a national overview of the variation in ED characteristics. Secondly, we investigated the association between ED volume characteristics (number of patients and hospitalizations) on one side and measures of available ED and hospital resources (ED treatment bays, ED nursing staff and number of hospital beds) on the other side. Obtained insight can be a starting point towards more well-founded future optimization policy.

This study, to our knowledge the first in its kind, demonstrates that the ED landscape in The Netherlands consists of a broad variety of EDs in regard to their individual profiles. In addition, on a national level this study identifies associations between patient and hospitalization volumes on one side and number of ED treatment bays, ED nurse workforce capacity and available hospital beds on the other side. These findings form a first step towards more insight into the ED landscape present and might be useful as input for the development of an ED resource allocation framework and more targeted optimization policy in the future.

Based on characteristics, we examined EDs in The Netherlands apparently cannot be considered as equals. In line with findings of previous research, these departments continue to show a pluriform landscape [4, 6, 14]. Reduction in the number of EDs over the years seems to have had little or no influence on this picture [6]. Other countries report comparable variation between ED profiles [19–22]. It is notable that with an average volume of about 23,000 patients (range 6000–53,000) and an average of 16 treatment bays (range 4–36) EDs in The Netherlands are relatively small in an international perspective. In 2016, the number of patients seen in an ED equals 116 patients per 1000 inhabitants. For comparison, in Belgium 290/1000, in the UK 264/1000 and in France 279/1000 inhabitants [32]. Together with both a low percentage of self-referrals (average 15%, range 1–56%) and a high percentage hospitalization of patients seen in the ED (average 38%, range 13–76%), in general EDs in The Netherlands might be considered as fairly efficient. At the same time, there is still a broad range of variation among EDs that is insufficiently understood. More insight into backgrounds of this variation and critical interpretation hereof might enable targeted interventions to achieve even more efficiency.

The second aim was to investigate relationships between a selection of volume and resource characteristics on a national level. Because there is no national ED framework yet, we hypothesized relationships as a first method of explorative research. No generally accepted relevant correlation hypotheses are available in literature. Therefore, we had to formulate them ourselves first. Despite limitations, the use of hypotheses can be seen as useful first indicators towards better understanding how available EDs relate to one another in regard to specific characteristics. We formulated six hypotheses around ED patient volumes and hospitalization of ED patients on one side and available resources on the other side and found both positive and significant correlations. Although this might seem obvious at first sight, these findings highlight at the same time the question that explains the differences present. How can it be that EDs with an equal volume of ED patients or hospitalizations do so with huge differences in terms of resources? These and other differences need to be investigated and explained in order to explore where there is actually room to improve quality and reduce costs.

In daily practice, our findings establish a baseline understanding of ED characteristics and mutual relationships. First of all, this can serve as a starter towards the development of an ED resource allocation framework and in line with this as input for an ED benchmark-instrument. This development however does require more comprehensive insight into characteristics of individual EDs, how these departments relate to one another and especially adequate explanations for mutual differences. For example, to compare ED resource planning, information such as distribution of patient volume and admission volume during the day/week/months, severity of the illnesses or injuries treated, and length of stay in the ED together with insight into reasons for admission should be taken into account. Our findings must be seen as a first step and give rise to a more extensive assessment of ED characteristics and interpretation of variation. Secondly, when coupled with mandatory reporting, annual assessment may become an instrument to determine targeted interventions for ED landscape optimization policy for reasons such as cost reduction, quality improvement and preventing crowding on one side and to monitor the effects of this policy in more detail on the other side.

Our study may also give direction to future research aiming to provide additional insight in the ED landscape and the effect of interventions on costs, quality and crowding. Thus far, we have used annual surveys on a voluntary basis. Further studies could benefit from a mandatory national ED registration coupled with more extensive reporting, including variables to correct for case mix differences on the level of patients (e.g., severity/complexity of patients) and ED units themselves (type of hospital, structure with GPC, geographical location). Future research is needed to achieve more and better insight into characteristics of individual EDs, the mutual connection of this, how individual EDs relate to national outcomes and adequate explanations for deviations. Studies are also needed to explore how, as far as ED care is concerned, economies of scale relate to scale disadvantages. In addition, it would be useful to identify objective national volume to resource ratios which can serve as a reference point for local interpretation. Our findings can also be of interest to the international reader. After all, a more detailed insight and understanding of the Dutch ED landscape makes it possible to compare more accurately with the ED landscape in their own country, for example when looking for optimization options.

Although our study has its strengths, like 100% participation and high degree of completeness of requested data, we acknowledge that several limitations may impact study results. The study was based on data obtained from registrations of individual EDs. We were not able to check for accuracy of all registries ourselves. Second, an (inter)nationally validated instrument to characterize EDs does not exist. In this explorative study, we investigated a selection of administrative characteristics. Although among others based on previously conducted and comparable research, we do not pretend to be complete or have used the only best method. Thirdly, as far as we have been able to verify, this is the first study using hypothesis to explore relations between characteristics within the ED landscape. No generally accepted relevant correlation hypotheses are available in literature. Therefore, we had to formulate them first. Although this is not validated, we believe this is a method worthwhile in order to obtain better insight. Despite the limitations, the use of hypotheses can be seen as useful first indicators towards better understanding how available EDs relate to one another in regard to their characteristics. Fourthly, we have studied centrally organized EDs that were operational and available 24 h a day, 7 days a week. Alternative hospital entrances for urgent care were not part of our study, but may be included in the future. Finally, the authors were aware that ideally comparing ED resource planning should take more basic information then patient volume and hospitalization volume into account. Unfortunately, this information was not available. Despite these limitations, our study provides more insight in the ED landscape of The Netherlands than was available to date.

In conclusion, our study shows that the ED landscape is still pluriform by numbers and specifications of individual ED locations. At the same time, this study identifies on a national level associations between patient and hospitalization volumes on one side and number of ED treatment bays, ED nurse workforce capacity and available hospital beds on the other side. These findings establish a baseline understanding of how EDs relate to one another and form a first step towards more insight into the ED landscape present and might be useful as input for the development of an ED resource allocation framework and more targeted optimization policy in the future.