If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Emergency Department, Wollongong Hospital, Illawarra Shoalhaven Local Health District, Wollongong, NSW, AustraliaResearch Central, Illawarra Shoalhaven Local Health District, NSW, AustraliaSusan Wakil School of Nursing and Midwifery, Faculty of Medicine and Health, RC Mills Building, The University of Sydney, NSW 2006, AustraliaGeorge Institute for Global Health, King St, Newtown, NSW, Australia
Increased Emergency Department length of stay impacts access to emergency care and is associated with increased patient morbidity, overcrowding, reduced patient and staff satisfaction. We sought to determine the contributing factors to increased length of stay in our mixed ED.
Methods
A real-time observational study was conducted at Wollongong Hospital over a continuous 72-h period. Times of intervention, assessment and treatment were recorded by dedicated emergency medical or nurse observers. The time from triage to each event was calculated and descriptive analyses performed. Free text comments were analysed using inductive content analysis.
Results
Data were collected on 381 of 389 eligible patients. The largest time delays were experienced by patients who required a CT, specialist review and/or an inpatient bed. Registrars and nurse practitioners were the most efficient in reaching a decision to admit or discharge. The time from triage to specialist review increased with the number requested (148 min for one, 224 min for two and 285 min for three). The longest length of stay was experienced by mental health and paediatric patients.
Conclusions
The main delays contributing to ED length of stay were CT imaging and specialist reviews. Overcrowding in ED need targeted, site-specific interventions.
]. Demand for emergency care and hospital admission is at an all-time high. This results in overcrowding. Overcrowding in EDs is defined as “where the capacity of a hospital’s inpatient services cannot meet patient demand, access block and emergency department overcrowding occurs, which the Australasian College for Emergency Medicine considers are critical indicators of health system dysfunction” [
Does volume or occupancy influence emergency access block? A multivariate time series analysis from a single emergency department in Sydney, Australia during the COVID-19 pandemic.
]. Other consequences include patients who do not wait for treatment, increased risk of iatrogenesis, higher operating costs and decreased patient satisfaction [
The causes of ED overcrowding are complex, multifactorial and can be grouped as follows. (1) Input: the volume and type of care presenting to ED, (2) Throughput: the internal delays associated with ED and (3) Output: the discharge of patients home or admission [
The volume, complexity and acuity of patients presenting to Australia’s EDs is increasing each year. Between 2017–18 and 2021–22 the total number of presentations assigned a triage category of Urgent or higher increased by 3% [
] and only 58% of urgent (cat 3) patients were seen within the recommended 30 min. In 2021–22, the proportion of ED visits completed within 4 h was 61%, down from 67% in 2020–21% and 71% in 2017–18 [
Throughput delays can be related to triage, room placement, initial provider evaluation, diagnostic testing and treatment while in ED. This includes cohesiveness of care teams, physical layout, nurse and physician staffing ratios, efficiency and use of diagnostic testing, accessibility to medical information, quality of documentation and communication and availability of specialty consultation [
Output delays are primarily related to access block which is the principle factor responsible for ED overcrowding and reflects hospital and health system performance rather than that of the ED [
]. For example, in a 2021 study, the main determinants in the reduction in ED overcrowding and access block were associated with reductions in hospital occupancy and elective surgery levels [
] in part by consuming ED nursing and physician resources.
With these considerations in mind, numerous initiatives have explored ways to mitigate overcrowding. The majority focused on ED productivity, rather than whole of hospital performance. Some effective ED efficiency solutions include nurse-initiated radiology and pathology [
Initiating diagnostic studies on patients with abdominal pain in the waiting room decreases time spent in an emergency department bed: a randomized controlled trial.
NEAT was implemented to reduce ED occupancy times. NEAT improved some internal ED efficiencies and remains a performance indicator (now called Emergency Treatment Performance - ETP) in NSW [
], and is impacted by access block. Aside from inpatient bed availability once a patient is admitted, data are not routinely reported that explain delays to a decision to admit, or why the patient may or may not have met the NEAT target. NEAT data are often incomplete, analysed retrospectively, and not truly reflective of ED functionality. For example, although many data are routinely and reliability time stamped and captured, such as time to CT request and report availability, other influential time points are not. For example, time of; blood collected, request for specialist team consult, medical assessment, analgesia. These times are either not collected, or, dependant on timely clinician data entry in the electronic medical record. To enable appropriate selection of interventions to improve ED throughput, sites should identify their own site-specific causes of delay. Our time in motion study sought to identify and quantify internal delays to enable selection of interventions most likely to result in improvement to ED throughput at Wollongong Hospital.
Methods
This real-time observational study was conducted from 16th April 0920hrs to 19th April 0700hrs, 2020 (Thursday – Saturday) at Wollongong Hospital Emergency Department (WH ED). This enabled 24/7 observation and was based on availability of additional staff to act in a dedicated data collection role. Wollongong hospital is the Illawarra Shoalhaven’s major referral hospital and the ED is one of the busiest in New South Wales with over 70,000 presentations annually. WH ED is a regional, mixed ED, that treats adults and children. From 2021/2022 WH ED registered 73,217 patients of which 68% started treatment within the recommended timeframe and only 47% left the ED within 4 h. The WH ED has 11 day, 11 evening, 6 night medical staff, 26 day, 26 evening, 23 night nursing and 5 day, 5 evening, 2 night clerical and 3 day, 3 evening, 2 night support staff on per day. The project was deemed a quality project (QA-109) and exempt from ethical review, complying with NSW Health Guidelines [
Human Research Ethics Committees. Quality improvement & ethical review: a practice guide for NSW (GL2007_020). Office for Health and Medical Research, 8; 2007.
The primary source of data was observation. Observation methodology is a systematic approach to analyse and improve work processes by carefully observing and recording the time and motion of workers. This methodology is useful to identify and eliminate inefficiencies, redundancies and wasted time in a process. A major advantage of the observation methodology is higher fidelity of information regarding specific tasks. This level of detail can help accurately identify inefficiencies and result in an overall improvement in productivity by generating more robust data [
]. Furthermore, observation methodology allows for the identification of potential sources of error or delay, that might not be captured through retrospective review of times for each item.
Ten ED staff were nominated by ED management, orientated to the data collection process, rostered to a supernumerary eight-hour day, afternoon or night shift. Each shift had two data collectors. ED staff were not formally advised of the study, however if staff asked, they were informed that the observers were collecting information about the department, and no identifiable information about any ED staff were being collected.
Patients were identified at triage by one of the data collectors, then tracked real-time through ED. The data collector took manual notes of every step of each patient’s ED journey using a template adapted from a Whole of Hospital patient flow project conducted at Wollongong in 2016 (see Supplementary material 1). They recorded patient details, date and time of event and any issues (in the comment section). The events that were recorded included time of triage, registration, first medical officer contact/review, test requests and availability (bloods, xray, ultrasound, CT), first specialist review, decision (admit/discharge), beds (requests, allocations, ready), transfer, admit/discharge and details of delays. The full list and further detail can be found in Supplementary material 1.
To verify time points secondary sources of data were used; pathology, radiology, staff recall and the electronic medical record. For example, to verify the time of xray, the timestamp from imaging was collected. To verify the time of blood sample collection, the time of pathology receipt in the laboratory was recorded. When the observer had not seen when and where a patient had been moved to, clinical staff were asked. When beds were allocated, and departure to that bed, the time entered by the clinical manager in eMR (Firstnet) was used, which, locally, is reliable.
Data management and analysis
Data were manually entered into Microsoft excel. The data underwent cleaning, double data entry, and were checked for missing data by three staff. Data were then de-identified. The time of 0 min was considered to be the timepoint when the patient was triaged. Time 0 min was triage. The patient timeline from triage to each event was calculated subtracting the time of triage from the event. Descriptive analyses were conducted and all results are presented in minutes. Normally distributed data are described with mean and standard deviation, skewed data are described with median and interquartile range. Kruskal–Wallis one-way analysis of variance was used to assess differences in time from triage to event across the five care providers: Intern, Registrar, Junior Medical Officer (JMO), Staff Specialists (SS) and Nurse Practitioner (NP). Post-hoc analysis was conducted using the Mann–Whitney U test. Results were Bonferroni corrected for multiple comparisons. The sample was not suitable for inferential statistics. All quantitative and descriptive analyses was completed using SPSS IBM v27. Free text comments were analysed using inductive content analysis. A content analysis was performed on the textual data based on a three phase the framework by Elo and Kyngas [
]. Data was read and reread by the researchers, condensed and coded, looking for frequency, similarity and differences. Categories were inductive, based on manifest content using a descriptive approach [
Data were collected on 97.9% (381/398) of patients who presented to WH ED between 0920hrs on the 16th of April and 0700hrs on the 19th of April 2020.
Demographics
During the 72-h study period 205 male and 193 female patients presented to the ED with a mean age of 41.2 years old (SD = 23.4). There were incomplete data on 17 patients, leaving 381 patients. The primary mode of arrival was private car (59%) and Ambulance (41%). Most patients were allocated a triage category three (44.8%) and four (34.7%), followed by category two (15.8%), five (4.1%) and one (0.6%). Almost two thirds of patients were discharged home (62%), followed by admitted to the ward (28.5%), critical care (5.7%), did not wait (3.3%) and transferred (0.5%).
The Patient Timeline – Triage The median time from triage to decision to admit or discharge was 151 min (IQR = 151 min). The largest time delays were experienced by patients who required a CT, specialist review and/or an inpatient bed (Fig. 1). The median time from triage to CT request was 61 min (IQR = 133 min), attending CT 163 min (IQR = 121.5 min) and a report available 215 min (IQR = 148 min).
Fig. 1Median and 80th percentile* time from triage to event in the ED.
*80th percentile demonstrated in line with local performance targets. MO = medical officer; NP = nurse practitioner; DTA = decision to admit; CT = computed tomography.
Nearly one half of the study sample were referred for specialty review (n = 171, 45%). The time spent in ED increased with the number of specialty reviews requested. If a specialist review was required, this referral occurred at 83 min (IQR = 126), and the review occurred 65 min later (148, IQR 178 min). If a second specialty review was needed this occurred at 224 min (IQR = 171.75), a third at 285 min (IQR = 262).
The greatest median time to decision to admit was for Obstetric and Gynaecology patients (228.5 min, IQR = 115.5), followed by mental health (183 min, IQR = 275) and paediatrics (178 min, IQR = 63.5). The shortest time was ICU 66.5 min (IQR = 117). This delay to decision making was reflected in the median ED length of stay for mental health (288.5 min) and paediatric patients (166 min). The time from bed allocation to availability varied by specialty of the admitting team. The longest times were observed for ICU (58 min, IQR = 151.5).
Relationship between time points and ED care provider
Significant differences were observed between the care providers from: triage to being seen by a medical or nursing practitioner [(MP/NP, H(4) = 10.89, p = .03; triage to specialist one review, H(4) = 10.14, p = .04; and, triage to decision to admit, H(4) = 27.46, p < .001]. NPs were significantly faster than interns in time from triage to review by the first specialist, U = 2.00, z = −3.01 (corrected for ties) p = .003]. Further, NPs were significantly faster than registrars, U = 1041.00, z = −4.14, p < .001, JMOs, U = 245.00, z = −5.02, p < .001, lead/SS, U = 325.50, z = −3.61, p < .001, and interns, U = 65.50, z = −4.47, p < .001, in reaching a decision to admit or discharge (Fig. 2).
Fig. 2Time from triage to event by treating staff type. Junior Medical Officer (JMO),.
Free text comments were initially grouped into 18 preliminary categories, then refined to eight categories and classified as throughput and output (Table 1). The most common throughput delays were waiting test results (for example repeat troponins or CTs), specialist team reviews and patient deterioration. Output delays were usually related to bed availability. These findings are reflective of the quantitative data.
Table 1Categories of delays through the ED.
Category
Exemplar
n
Through-put
Patient factors
Patient deteriorated in ED, patient not being between the flags
8
Emergency delays
ED treatment not completed before bed available / Delay in MO seeing patient after picked up from the list / no staff to transfer patients
4
Waiting test results
Labs results like serial troponin or delay in imaging
27
Inpatient team delay
Difficulty in inpatient team admission/ Awaiting specialist opinion / Delay in med reg review
14
Communication issues in ED
NUM not told of admission
8
Total
56
Output
Ward delay
Ward not ready/ no bed available/ Covid delays i.e. ward not accepting until swab done
16
Transport delay
Awaiting transport for discharge/ awaiting hospital transfer/ awaiting discharge letter
10
Change in disposition
Admission vs discharge or change in admission destination
This project tracked patients in real-time through the ED to identify factors contributing to ED throughput. The most significant causes of delays were CT, specialist review and bed availability. Other common delays were waiting to be seen by a health practitioner, ward transfer and specialty review which escalated exponentially with more than one review.
Our findings are concordant with existing literature that demonstrated ED length of stay is multi-factorial [
]. In our study, awaiting medical imaging and reporting was the cause of the longest delays. This is could simply be a mismatch between supply and demands. We know that local demand for Emergency CT is increasing 10% year on year. We cannot comment on the effect of over investigation and unnecessary imaging. It is known that early ordering of imaging by senior doctor after a rapid assessment, reduces the ordering of unnecessary CTs [
Initiating diagnostic studies on patients with abdominal pain in the waiting room decreases time spent in an emergency department bed: a randomized controlled trial.
]. In contrast to medical imaging delays, the time to blood collection and results in our study was relatively short. This could be explained by an efficient pathology service, embedding of nurse-initiated blood tests [
Another major contributor to length of stay was referral to specialty teams. This is consistent with Bashkin et al. three month observational study of 105 patients which demonstrated communication between ED and inpatient teams caused significant delays in admission time [
]. Patients have increasing complex chronic medical conditions and therapies. When more than one specialty review was required, delay increased exponentially. We found variation between specialties which may reflect junior doctor availability for ED consults and other responsibilities when on-call and senior specialty support for decision making. Admission policies that contain clear criteria and processes for specific clinical conditions would assist with selecting the appropriate team initially and holding that team accountable to accept admissions.
Delays in the availability of a ward bed also contributed to increased ED length of stay. This issue is multifactorial with delays in patient discharge due to lack of residential aged care or community support, lack of ward beds and ward staff [
]. Since this study was concluded, our health district has lost residential aged care beds, resulting in up to 160 inpatients awaiting community placement at one anyone time [
]. Since the conduct of this study, the NSW Government commissioned a Parliamentary Enquiry to ED overcrowding and the Australasian College of Emergency Medicine (ACEM) has released a position statement outlining possible solutions informed by a 2022 review by the Sax Institute [
]. These include: a whole of system approach comprising system wide change; increasing hospital and alternative health care, matching discharges with daily demand, having over capacity protocols shared throughout the hospital, having time based targets throughout the hospital and extending hospital function to 24/7; reducing hospital inpatient bed demand through enhancing hospital in the home, ambulatory care clinics, community based mental health; and creating an evidence base to inform policy development.
Although our study was a unique in that it captured patient flow in real time as well as commentary, there were some limitations. Data were collected during the first wave of COVID-19, when presentations and in hospital activity was reduced. Should this be conducted currently, we anticipate the findings to be exacerbated. Alternatively, the number of admissions in some specialties was quite small, and data skew may not be totally accounted for in use of median (IQR), resulting in inflated findings. Our findings inform local processes, and extrapolation should be made with caution. A larger sample size across multiple sites would add rigour.
Conclusion
The main delays contributing to delays in ED throughout were imaging and specialist reviews. Any efficiency interventions for ED need evidence informed, site-specific interventions. ED throughput is just one aspect of ED overcrowding. For real change, a whole of hospital whole of system approach which including increased alternative health care options capacity, having over capacity protocols shared throughout the hospital, and extending hospital function to 24/7.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Ethics approval
This project was approved as a quality activity by the Illawarra Shoalhaven Local Health District Low and Negligible Risk (LNR) Research Review Committee (REF 411).
Disclosure statement
The authors whose names are listed certify that they have NO affiliations with or involvement in any organisation or entity with any financial interest, or nonfinancial interest in the subject matter or materials discussed in this manuscript.
Does volume or occupancy influence emergency access block? A multivariate time series analysis from a single emergency department in Sydney, Australia during the COVID-19 pandemic.
Initiating diagnostic studies on patients with abdominal pain in the waiting room decreases time spent in an emergency department bed: a randomized controlled trial.
Human Research Ethics Committees. Quality improvement & ethical review: a practice guide for NSW (GL2007_020). Office for Health and Medical Research, 8; 2007.
00027-1&id=gr1.jpg){kind=link}
00027-1&id=gr2.jpg){kind=link}