Introduction
Increased length of stay (LOS) in crowded emergency departments (EDs) is a global problem [
1,
2] that is associated with reduced quality of care [
3‐
6]. A variety of techniques, such as team-based triage, fast-tracking, laboratory analysis in EDs, and nurse-requested X-ray imaging have been suggested [
7]. Time targeted policies have also been proposed and implemented in several other countries, including the UK, Australia, New Zealand, and Canada [
7‐
17].
In response to the Middle East Respiratory Syndrome (MERS) epidemic of 2015, South Korea introduced a “24-hour target policy” for EDs to prevent crowding and reduce the average LOS [
18]. At the time of the South Korean MERS pandemic, among a total of 186 patients, 82 people were infected in one crowded ED due to a super-spreading event from one patient [
19,
20]. This has led to a social consensus that ED crowding should be addressed. As a result, the 24-hour target policy was introduced in December 2017.
The 24-hour target policy requires at least 95% of all patients to be admitted, discharged, or transferred from the ED within 24 hours of arrival. Although there is a crowding disparity between EDs in Korea, mean proportion of patients who stayed in the ED for more than 24 hours reached 10% in crowded EDs [
21,
22]. This characteristic crowding of Korean EDs led to a policy target time of 24 hours, which is longer than that used in other countries.
Setting targets might increase organizational performance; however, target-driven care risks distorting clinical priorities [
23]. In previous studies, time targets in EDs have yielded controversial results, with both positive and negative consequences beyond their intended effects [
7,
14‐
16,
24]. The effects of the South Korean policy on patients and medical professionals have yet to be studied and, given the longer target time of 24 hours compared to policies implemented in other countries, differences in impacts can be expected.
This study aimed to identify the impact of the 24-hour target policy in Korea on patients and medical professionals. This study used a mixed-methods approach to study the impact of the policy on LOS in EDs, policy compliance rates, and other consequences for patients and medical professionals.
Methods
Overall approach
The study is a retrospective observational study using a mixed-methods design to analyze ED visits and a survey of ED medical professional experiences. This study was conducted at a tertiary referral hospital in South Korea.
This study was approved by the institutional review board (IRB) of Samsung Medical Center. The need for informed consent was waived due to the retrospective, observational, and anonymous nature of the study by the institutional review board (IRB) of Samsung Medical Center. (IRB No. 2021–08-172). The survey of ED medical professionals was approved separately with informed consent (IRB No. 2021–08-173) of Samsung Medical Center.
Participants and data sources
This retrospective study was conducted in the ED of a tertiary metropolitan hospital with approximately 1960 inpatient beds and approximately 80,000 ED visits per year. This study included ED visits from February 1, 2016, to June 31, 2019. The plan for implementing the 24-hour target policy was announced on July 10, 2017, and the policy was implemented on December 3, 2017 [
18]. We classified the research period into three parts: pre-policy (February 2016 to June 2017), adjustment period (July 2017 to January 2018), and post-policy (February 2018 to June 2019). We compared outcome measures between the pre-policy and post-policy periods to evaluate the impact of the policy.
Survey responses were collected from 22 doctors and 39 nurses over 2 weeks, from November 3 to November 17, 2021. All respondents worked in the hospital both before and after implementation. The mobile questionnaires were filled out using Google.
Outcome measures
The primary measure was ED LOS, and the secondary measure was the policy compliance rate. The time target policy requires 95% of patients to be admitted, discharged, or transferred from the ED within 24 hours of arrival. The policy compliance rate refers to the proportion of patients who successfully moved out from the ED within 24 hours. Along with ED LOS, the proportion of patients remaining in the ED after 24 hours was used as an indicator of the policy application.
Tertiary measures included the following outcomes: time to first prescription, time to admission decision, time to admission, time to computed tomography (CT), time to magnetic resonance imaging (MRI), time to operation, time to coronary angiography (CAG), and proportion of patient dispositions determined at 23 hours. All time variables, except time to admission, were calculated from the first presentation; time to admission was calculated from the time the decision to admit the patient was made.
Survey
Patient flow, quality of care, patient safety, staff workload, need for improvement of the policy, and staff satisfaction levels were investigated through a questionnaire. Patient flow included the overall, triage, diagnostic evaluation and treatment, and disposition process. Quality of care was also assessed in terms of patient-centered, safe, effective, timely, efficient, and equitable treatment. In addition to the overall safety component in the quality of care section, questions for patient safety including patient identification, pressure ulcers, falls, medication, diagnostic tests, treatment, and others (infection-related, medical equipment, escape, violence, blood transfusion, etc.) were included. Workload dimensions, including mental, physical, and temporal demand, performance, effort, and degree of frustration, were also assessed. On top of the satisfaction of mefical staff with the policy, the degree to which medical staff felt the need for improvement in each patient flow was investigated. To develop a more comprehensive understanding of staff experiences, we included open-ended survey questions about aspects of the ED experience related to the time target policy (Supplementary table
1).
Questions about patient flow were written based on the input-throughput-output conceptual model of ED crowding suggested by Asplin et al. [
4]. Questions about quality of care were based on the six domains of quality of care established by the Institute of Medicine and Medical Office Survey on Patient Safety Culture of the Agency for Healthcare Research and Quality. Questions on patient safety were based on the Korean Patient Safety Incident Report 2020 by the Korea Institute for Healthcare Accreditation [
25,
26]. Questions about workload were based on the National Aeronautics and Space Administration Task Load Index [
27].
Statistical analysis
Continuous variables are presented as medians and interquartile ranges (IQRs) according to non-normal distributions on the Anderson-Darling test. Categorical variables are expressed as frequencies and percentages. To compare patient visits before and after policy implementation, the Mann–Whitney U test was used for continuous variables that were not normally distributed, and the χ2 test was used for categorical variables. P-values < .05 were considered statistically significant. All statistical analyses were performed using R (version 4.1.1; R Foundation for Statistical Computing, Vienna, Austria).
Discussion
To the best of our knowledge, this study is the first mixed-methods analysis of the impact of the 24-hour time target policy on ED experiences in Korea. The time target policy for EDs was introduced with the expectation of a whole-system approach to improving ED LOS [
12]. This study found an increase in overall ED LOS and time to some ED processes despite good policy compliance.
In South Korea, patients can freely visit tertiary hospitals even when they don’t have a referral from a primary or secondary provider [
28,
29]. This is unique among national healthcare systems, and the trend toward increasing demand at several already-crowded tertiary hospitals is intensifying because of the recent decision to strengthen health insurance coverage [
30,
31]. Moreover, many South Korean patients are awaiting diagnosis and treatment for complicated chronic diseases or hospitalization for continued treatment following acute treatment in EDs [
30]. The resulting crowding of EDs at tertiary hospitals led to the introduction of a 24-hour time target.
After implementing the time target policy, the proportion of patients with an LOS exceeding 24 hours decreased significantly, although the median ED LOS increased slightly (Figs.
1 and
3). Despite achieving the surface goal of reducing the proportion of patient stays exceeding 24 hours, the time target policy did not reduce LOS or improve the overall ED flow which were the policy’s ultimate goals. Even considering the possibility of patient severity differences between the two periods, Supplementary table
2 shows that LOS increased in all KTAS groups except KTAS 1. The increase in ED LOS can be attributed to a significant increase in the number of patient visits after policy implementation. Among the three flows described by Asplin [
4], as input flow increases, more efforts is required to improve throughput and output flow. In the survey results, we observed similar complex responses; 63.9% of medical staff reported that patient overall flow seems to be improved or very improved (Supplementary table
4), and 47% of respondents were satisfied with the policy while 36% were not.
In this study, ED LOS distribution for 23 hours, just before the time target, changed before and after the policy. This finding is consistent with earlier research that found that, when a time target policy was implemented in other nations, including the UK, Australia, and New Zealand, an ED patient’s disposition tended to be determined just before the time target was reached [
24,
32,
33]. As a result, it is unclear whether the decrease in the proportion of patients staying in the ED for more than 24 hours was due to improved patient flow or to “gaming” the system by seeming to comply with the policy, as suggested by Tenbensel et al [
24]. Some survey participants in this study also described such “gaming” practices, along with the forceful transfer and discharge of patients. They also noted that urgent transfers or discharges could threaten patient safety owing to treatment discontinuity and insufficient medical staff in wards.
Despite the policy implementation, the time to critical tests and interventions which are classified as the throughput flow of Asplin [
4], increased further. The change in patient distribution clearly demonstrated an increase in the number of patients who needed tests and interventions during the same period (Supplementary fig.
1). Therefore, given the increased time to individual examinations and interventions, the distribution of ED LOS of 23 hours, and “game” effect, only the superficial goal of a 5%, the proportion of patients who stayed more than 24 hours, was achieved and it is likely that the time target policy didn’t work for improving LOS or the overall ED flow. It may be difficult to achieve the policy’s goal without controlling the input flow in Korea’s national health system, where health insurance coverage continues to increase and patients have a wide range of hospital options and a strong preference for tertiary referral hospitals.
A previous study in New Zealand showed ED LOS monitoring strategies including the display of real-time information for ED LOS and the operation of short-term emergency wards that only admit patients from the ED can help lower ED LOS and improve patient flow [
24]. Considering that the policy compliance rate was improved in the several months following the introduction of the time target policy, it can be expected that additional efforts were made to improve throughput and output flow. The operation of emergency wards that admit patients only from the ED, a dedicated transfer-coordinator nurse system for EDs, LOS management implemented by each department, and LOS monitoring within the ED would have been helpful in effectively managing policy compliance rate in target hospitals [
34‐
37]. As such, if it is difficult to improve the input flow in the Korean healthcare system, looking at other throughput and output flows can be an alternative.
Meanwhile, unless there is a change in other conditions such as the number of medical staff, the fact that the timing of first prescriptions and tests are similar with the increased number of patient visits suggests that the burden on medical staff might have increased. The survey also identified increased workload among medical professionals. Medical staff reported that patient flow, quality of care, and some patient safety indicators were improved by the policy, but the workload of the medical professionals was greater than before the policy. Overall, the higher compliance rate despite the increased input flow represents the increased workloads of medical staff, which was presented in the survey. To pursue the target, the pressure to discharge a patient within 24 hours may also influence staff workload. Whether it is a “gaming” effect or an improvement in patient flow, the fact that the disposition of many patients is hastily decided at the 23rd hour compared to before the policy seems to have caused additional workload for the medical staff (Fig.
2).
The downside of this policy is that the medical conditions of patients are not considered. Patients who require additional workup or emergency care can be admitted to the ward or transferred to comply with the policy. Patients who require hospitalization often require more treatment time than patients who return home, as noted in previous studies [
32,
38]. In Canada, the target time varies depending on the severity of the disease or trauma and the acuity level of the triage stage; this approach could be applied in South Korea [
14]. In our survey, medical staff also expressed concerns about unified policy applications. Some medical personnel noted due to the pressure to comply with the time target, time and opportunities to take care of patients are insufficient.
Our findings suggest several approaches for improving the implementation of the time target policy. First, such policies require support from outside the ED, including an increased ward capacity and improved transfer systems. Second, each flow of ED, including triage, prescription, lab test, imaging tests, admission or discharge should be monitored and backed up to improve the entire ED process. Third, the workload of medical staff should be considered, and appropriate compensation should be offered when policy compliance is high. Fourth, disease severity of patients should be considered. Patients who visit tertiary hospitals have relatively high disease severity, and in many cases, it is difficult to transfer them to other hospitals. For patient safety, it may be helpful to adjust the target time according to patients’ severity, such as KTAS level or disposition, rather than applying a unified time target to all patients. Fifth, policies should be tailored based on the characteristics of each hospital. Each hospital has different patient characteristics and resources. In Korea, the severity and number of patients vary according to the ED location. Usually, patients with ED in metropolitan areas have higher severity and larger numbers. In tertiary referral hospitals located in metropolitan areas, patients often come from other tertiary general hospitals to receive appropriate treatment and additional medical resources. Policymakers should consider these factors to improve patient flow in Korea.
Limitations
This study has some limitations. As this was a single-center study, we were unable to exclude selection bias, and our results might not be generalizable. However, various participants and circumstances can be recruited from the study site, which is an crowded tertiary referral hospital located in an urban area. This study attempted to show several aspects of the implementation of the target policy such as the ED LOS, the time change of each process, and the ED LOS distribution. In addition, a survey was conducted on several aspects, including opinions on patient flow, quality care, patient safety, workload, and the process that needs improvement. Hospitals in similar environments can obtain a rich perspective from this study. Second, retrospective surveys face the risk of recall bias. An in-depth interview study might be required to improve quality assessment. Thirdly, the large study population must be considered when assessing the study’s results and the
p-value [
39]. The interquartile range and sample size were all expressed for additional interpretation. In addition, the time difference before and after the policy shownin Table
2can be clinically meaningful, even if it only 0.5 hours, and can influence all of LOS, overall ED process, and patient outcomes. Fourthly, Due to potential confounding factors, the interrupted time series analysis was not performed. Trends are susceptible to change at various times, especially during certain seasons. Instead, the research period was divided into three parts: pre-policy (February 2016 to June 2017), adjustment period (July 2017 to January 2018), and post-policy (February 2018 to June 2019), with concurrent months separating the pre-policy and post-policy groups. Finally, during the study period, other actions to improve ED flow were implemented or were already in place that could influence the ED LOS. The Ministry of Health and Welfare conducts annual quality evaluation through indicators such as the ED LOS of patients with severe ill code and the proportion of severe patients with high KTAS levels who were directly evaluated in a timely manner by emergency medicine specialists [
18,
40]. This was another reason that interrupted time series analysis was hard to perform. It was impossible to consider all these factors simultaneously because of the retrospective observational nature of the study. Instead, we focused on analyzing the Korean 24-hour target policy from a variety of perspectives, including both empirical data and surveys.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.