Patients Left Without Being Seen in the Emergency Department – A Retrospective Observational Patient Record Study

The share of patients leaving the emergency department without being seen by a physician (LWBS) is an important ecacy marker of emergency departments around the world. It has not however, been of special research interest in the Nordic countries. In this study, we assessed the LWBS rate in Kanta-Häme Central Hospital, in Hämeenlinna, Finland. Secondarily, we gathered information of LWBS patients along with all ED patients.


Conclusions
The LWBS rate in the Kanta-Häme Central Hospital ED was low. Higher LWBS rate coincided with rush hours and longer waiting times. The proportion of minor traumas was greater than in some earlier studies. These results support the implementation of a fast track line for minor traumas in the ED.

Background
The share of patients leaving the emergency department (ED) without being seen by a physician (LWBS) is considered one of the main e cacy markers of EDs worldwide [1,2]. It has not, however, been of special research interest in the Nordic countries. The main reasons for LWBS are overcrowding in the ED and long waiting times [1,3,4]. LWBS patients are considered a consequence of overcrowded or poorly functioning EDs and may ultimately, although rarely, lead to serious harm to such patients [5].
Our primary aim in this study was to assess the LWBS rate of the Kanta-Häme Central Hospital (KHCH), a middle-sized secondary teaching hospital in Finland. The secondary aims were to study the pro les of LWBS patients together with seasonal and temporal variations. Further, we attempted to identify factors which, if remedied, would help us to improve the e cacy of the ED. This study is a part of the research and reform projects conducted to systematically improve the processes in KHCH ED, such as the reorganization of our stroke treatment protocol [6,7,8,9,10].

Methods
The data consist of all patients attending the ED of KHCH over a 12-month period (n = 41,631). We evaluated LWBS patients from patient records. This identi cation was made from the so-called E-book, where patients checked in by a triage nurse and leaving without being seen by a doctor are registered. In our ED, every patient entering is triaged. Reasons for attending were collected from patient records as written by the triage nurse. We divided these into 14 categories based on the study by Carron et al. [4].
The date, day of the week, time of arrival, time of departure, age, gender, possible return to our hospital or other health care unit within seven days and 7 to 30-day mortality were also collected for statistical analysis. No triage category was obtainable retrospectively during the time of data collection.
The absolute number and rate of LWBS patients with seasonal and temporal variations were calculated.
Further, all patient visits to the ED and their absolute waiting times per hour together with mean and median waiting times with 95% con dence intervals were collected.
The study was a retrospective patient record study thus requiring no consent from patients. No ethical statement on the study was required the ethics committee of Tampere University Hospital according to

Results
We identi ed a total of 305 LWBS visits, 99.0% of whose arrival times were available (n = 302). The 305 visits consisted of 299 individual patients, six patients made two separate visits ending in LWBS. Time of departure was available for 55.4% (n = 169) of visits and reason for attending for 99.3% (n = 303). The characteristics of LWBS patients compared to those of all ED visits are shown in Table 1. Of these patients, 28.5% were less than 18 years of age. When comparing LWBS patients to the normal age distribution of all our patients, we can see an overrepresentation of age groups ≤ 35 years of age and male gender. Distributions of LWBS patients are presented as numbers (N) and percentages (%) according to gender and age groups. LWBS rate (%) in different age groups and genders is shown in the nal column. ED = Emergency Department; LWBS = Left Without Being Seen.
The overall LWBS rate was 0.73%. A signi cant monthly uctuation in LWBS rate was seen peaking in June at 1.40% (Fig. 1).
Most LWBS took place during daytime. The number LWBS patients re ected well with the total number of patients in the ED (Fig. 2a). The waiting time (n = 166) before LWBS was 75.5 minutes (median) and 89.5 minutes (mean; 95% con dence interval 79.3-99.7 minutes). For the whole patient population, the average mean waiting time before seeing the doctor varied between 48 and 92 minutes according to the time of the day, the shortest being at nighttime and the longest during the afternoon hours (Fig. 2b). A slight increase in waiting time to doctor can be seen at seven o'clock in the morning, which is most likely attributed to shift changes of doctors and nurses between seven and eight o'clock.
The most common cause for attendance ending in LWBS was minor trauma (29.4%) and the second most common cause was non-speci c pain (12.9%) ( Table 2). Adverse effects of alcohol and/or drugs were the main reasons for attending in only 3.6 % of LWBS cases but had some effect on a larger proportion of cases. There were only 214 patients whose native language was not Finnish or Swedish, the native languages in Finland (0.5% of all ED visitors). Their share in the LWBS group did not differ from that in the total population of the catchment area.  19,20]. Rush hours and longer waiting times coincided with larger numbers of LWBS, even though the total rate was low. Younger patients and patients with minor traumas formed the largest LWBS groups in our study.
Waiting times and rush in the ED are well studied factors associated with increased LWBS rates worldwide [4,14,16,18,19,20,21,22]. In this study we noticed that when waiting times before being seen by a physician exceeded 75 minutes, mostly during the afternoon hours with more patients, the LWBS rate increased. In some studies, nighttime was a risk factor for LWBS, but this was not apparent in the present study [15,17,23].
In our study hospital the proportion of minor traumas among LWBS patients as the reason for attending was rather high ( [15,16]. Younger age groups with a slight male preponderance were over-represented in LWBS in our study, which has also been reported in other studies [14,15,24].
In some studies, non-native language speaking background or non-compensable coverage status have been a strong predictor for LWBS [25,26,27]. In this study, however, patients whose native language was not Finnish were not overrepresented in LWBS. In Finland, all patients are covered by a municipal tax and the visit fees of EDs are low. Due to the retrospective nature of this study, visit fees as a cause for LWBS were not speci cally elicited. However, according to the feedback query system of our ED, visit fees have been a cause of complaints only in sporadic cases. Symptom relief was probably also in some of our patients associated with LWBS as also in some earlier studies [28,29].
A strength in this study was the large number of total patients studied. We believe that LWBS patients were comprehensively included in the study, since our health records (including the E-book) are well kept and also linked to the municipal payment transactions. The reason for attendance could also be reliably ascertained since on entering the ED every patient was seen by an experienced triage nurse. This yields a trustworthy estimate of the true proportions of different causes for LWBS patients attending the ED. Further, mortality is based on national records in Finland, which are comprehensive and reliable [30].
The limitations of this single centre study are mainly due to the study design, as it is retrospective. In a prospective study, we would have been able to specify the reasons behind LWBS and waiting times even more accurately than now. On the other hand, patients admitted to a prospective study might be less inclined to leave before seeing a doctor, thus causing a bias. Not all health records in Finland have so far been fully connected to the national database. Therefore, we have no means of knowing about every visit made to private sector healthcare facilities.
Reducing the LWBS rate is an important factor in present-day goal-directed ED management. This improves both patient satisfaction and safety as, although complications are rare, these may occur.
Applying lean healthcare interventions or speci c fast-track lines for patients improves LWBS rates and throughput in the ED [31,32,33]. These suggestions may further improve our results, since the proportion of children and minor traumas suitable for the fast tract policy was high in our LWBS population. To tackle the LWBS problem, we have already implemented an "ED lobby project" aiming to reduce waiting times for ambulatory patients and applying the principles outlined earlier [13].

Conclusions
The LWBS rate in our ED was low, only 0.73 %. The reasons for attending among LWBS patients differed slightly from those reported in other studies as minor traumas were over-represented. Our results support the implementation of a fast-track line for minor traumas, which could further improve the LWBS rate.
Since waiting times to physician contact were more than 75 minutes, especially during the afternoon hours, adequate targeting of resources could further help to reduce the LWBS rate.

Availability of data and material
Coded data is available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This study was funded by the Ministry of Social Affairs and Health through the research funds of Kanta-Häme Central Hospital and Tampere University Hospital as well as a grant from the Finnish Society of Emergency Medicine.
Authors' contributions AP and TI planned the study. MH analyzed the data with help of AP and TL. MH drafted the manuscript together with AP. AP was the responsible investigator. TI gathered the original data from the medical records with the help of TL. VH participated in analysis and drafting the manuscript. Finally, all authors read and approved the manuscript.

Figure 1
Monthly rate in LWBS patients of all ED patients as percentages.