The impact of the supervision structure on Emergency Department flow.

(1)

The impact of the supervision structure on

Emergency Department flow.

30-01-2017

A case study on the ED of the Martini Hospital

MSc Thesis Supply Chain Management

University of Groningen

Faculty Economics and Business

Luc Hoeben

s2058138

l.h.a.hoeben@student.rug.nl

Eeldersingel 46E

9726 AS Groningen

(+31)625431863

(2)

Abstract

Emergency department (ED) crowding reduces patient flow, adding non-value added waiting time to the length of stay (LoS) of patients. In Dutch hospitals, emergency care is provided by residents under the supervision of specialists to ensure the quality of the provided care as well as for educational aspects. This thesis clarifies how the supervision structure influences ED patient flow. An embedded case study is conducted on the Surgical and Internal Medicine specialty in the ED of the Martini Hospital, located in Groningen, the Netherlands. By performing quantitative and qualitative research methods, this research concludes that via supervised early assessment of the patient’s trajectory and protocols on the way supervision has to be provided regarding timing of supervision, patient flow can be optimized by reducing non value-added waiting time and improving the resident’s logistic responsibility.

Keywords: crowding, emergency department (ED), patient flow, supervision structure, resident,

(3)

1. Introduction

On the 6th of June in 2016, the Minister of Health, Welfare and Sport of the Netherlands initialized a nationwide research on why emergency departments (ED) initiate patient stops on a daily basis due to overcrowding and temporarily refuse to admit new patients as a result (Schippers, 2016). This research is based on a letter written by the nation’s acute care committee, stating the problems of crowding in Dutch emergency departments nationwide (ROAZ, 2016).

Downside of ED crowding is the relation between crowding and poor quality of healthcare and declining patient safety (Carter, Pouch & Larson, 2014). Crowding distorts the ideal flow throughout an ED, explained by the American College of Emergency Physicians (2006), stating that crowding involves: “A situation in which the identified needs for emergency service outstrips available resources in the emergency department, hospital or both”. This definition comprehends the term ‘resources’, which is an important asset in the existence of crowding. Resources is a widespread term which in a hospital setting can be seen as the amount of staff, beds and equipment that is available for healthcare (Moss et al., 2014). When these resources are not available, non-value adding waiting time will be inflicted upon the regular process time, and thereby reduce the quality of the treatment (Joosten, Bongers & Janssen, 2009; Carter et al., 2014) and distort the patient flow (Drupsteen, 2013).

(5)

guidance and feedback on matters of personal, professional and educational development in the context of the resident’s care of patients” (Kilminster & Jolly, 2000). The supervision structure exists of multiple components. In general, this structure involves the delegation of responsibilities and tasks from the supervisor towards the resident determining how autonomous residents work. A part of the structure is explained as the way supervision is given and via what medium e.g. face-to-face or by phone. At last, the timing of supervision provision in the patient process plays a role in this structure.

The goal of this research is to identify and provide insight in how the supervision structure contributes to optimal patient flow. The focus of this thesis will lie on the following research question:

How does the supervision structure influence the flow of an emergency department?

To assess the impact of this structure, a case study is conducted within the ED of a peripheral hospital in the Netherlands. An intervention is implemented where changes in the supervision structure are made with the goal to improve patient flow. A before and after study determines the impact of the intervention and is aimed to explain how supervision can be used to affect ED flow.

(6)

2. Theoretical Background

Within the theoretical background, the basis for this thesis will be grounded in literature, providing a framework of previous research and address the importance of this study. Multiple phenomena will be clarified such as crowding and supervision, and will provide a structured framework for the methodology of this thesis.

2.1 Crowding

When patients enter the emergency department faster than there is capacity to treat them, crowding occurs (Drupsteen et al., 2013). ED crowding leads to extensive throughput times. Even in a non-crowded emergency setting, LoS is directly related to the quality of the provided care (Thomas, Guire & Horvat, 1997; Pines, Garson & Baxt, 2007), which explains why much research is dedicated to improve the LoS of patients. Where research has been devoted trying to explain how an ED operates and why it is such a complex environment, it is still hard to grasp what the solution is that fixes crowding. This mainly has to do with the unique characteristics of every hospital, staff, patient, country, political system and region (Hoot & Aronsky, 2008). An up-to-date structured review for literature regarding crowding is provided by Jha, Sahay, and Charan (2016), explaining crowding as a widely studied phenomenon where no generalizable solution can be provided for due to its multicomponent nature.

Although there is no explicit data about emergency department crowding in the Netherlands and relevant literature about the Dutch system is scarce, crowding happens on a daily basis in Dutch hospitals all around the country (Van der Linden, 2013). In comparison to the UK and the US, the impact and volume is much less, but it is a phenomenon with negative impact throughout the country accounting for lower quality of care (ROAZ, 2016).

Although generalizable solutions are still unavailable, Derlet & Richards (2000) investigated the causes of crowding, and identified fourteen major causes. Three of these causes have identified the importance of adequate supervision within the ED, due to the required knowledge and skillset of specialists and residents:

- Avoiding inpatient hospital admission by “intensive therapy” in the ED - Increased complexity and acuity of patients presenting to the ED - Shortage of on-call specialty consultants or lack of availability

(7)

2.2 Supervision

Kilminster and Jolly (2000) provide a review basis for literature on supervision. They conclude that research is limited and is mostly theoretical due to its complex nature. Where research is done on the educational aspects of supervision, no generalizable link between supervision and quality of care can be provided. This is supported by the research of Snowdon, Hau, Leggat and Taylor (2016), who argue that supervision can be linked to a reduction of patients’ mortality risk and overall complications, although the evidence for this conclusion is low and may not be applicable for the entire industry of healthcare.

To understand the multicomponent nature of supervision, Loganbill, Hardy and Delworth (1982) identify multiple facets that determine how supervision is constructed. They state that characteristics of the supervisor, the supervisee, the relationship between them, the environment and the client determine the degree, amount, and sort of supervision. These elements need to be assessed and dealt with by the supervisor to determine the necessary action. Their model shows that supervision in a medical setting, also called clinical supervision, requires a tailor made approach, and no generalizable standard exists for optimal supervision to provide high quality care.

Ladany, Mori and Mehr (2013) argue that of the characteristics of Loganbill et al. (1982), the relationship between the supervisor and supervisee is the single most important factor for effectiveness of supervision, explaining the difficulties in setting a standardized setting for optimal supervision due to the uniqueness of every human being.

Where the research of Loganbill et al. (1982) addresses supervision in a general medical context, it lacks a more in depth view about how supervision is constructed in different situations. Delaney, Zimmerman, Strout and Fix (2013) explain in their research that next to a caring and curing facility, a hospital is an education platform. Practice makes perfect and so routine and knowledge has to be gained in real life situations. Concluding, while giving care as a resident and specialist, there is a tradeoff between education and having autonomy on caring for a patient. This shows that within the provision of supervision, educational aspects play a role while providing the highest quality of care (DeLaney, Zimmerman, Strout & Fix, 2013).

(8)

supervision of the supervisee’s performance. These factors address the importance of acute

presence of the supervisor and show the value of reducing waiting time for supervision. The need to transform ad-hoc decision forming to planned due to lack of time when acute responses to an event are needed for an optimal supervision structure.

2.3 Supervision within the ED process.

Planned decision forming by ED staff relates to logistic decision forming regarding patient flow. These choices determine the trajectory of the patient throughout the ED and hospital, and let the clinician have control over the system (Kingsman, Rotter, James, Snow & Willis 2010). Not having this control leads to inefficient decision forming and leads to crowding, affecting the pathways of other ED patients and inpatient wards in the hospital (Tang, Chen & Lee, 2015).This control can be explained as the overview of the requirements for the patient needs in terms of resources they need and the physical location the patients will be within the process.

Nugus (2011) discusses the importance of logistic decisions as they are of equal worth as the diagnosis itself. Examples of logistic decisions are the need for hospitalization and determining what tests (e.g. radiography, liquid) the patient requires. Early assessment of the logistic situation in a process creates a responsive manner to cope with discrepancies, creates a more robust system and improves flexibility to reduce working pressure by managing scarce time and space in the ED (Nugus, 2011). It actively constructs a flow throughout the ED, instead of flow appearing as a consequence of efficient handling. For instance, much research has determined the delay in patient transport of hospitalized patients as the prime reason why crowding occurs (Derlet, Richards & Kravitz, 2001), explaining the need to assess this as early on in the process if possible to plan these movements ahead.

(9)

Relating supervision to crowding in emergency care, crowding leads to less compliance of clinical guidelines and inappropriate decision making (Miro et al., 2003). More compliance with guidelines due to direct supervision comes forward as one of the primary outcomes when it comes to a supervisor-resident relationship (Kingsbury & Allssop, 1994). Combining these two statements, less compliance of clinical guidelines can be prevented by early interventions and so, early supervision could optimize the quality of care in a (crowded) ED.

2.4 Supervision Structure

Although no literature has linked the structure of supervision to patient flow, Kilroy (2006) speaks about an effective clinical supervision structure in an ED with the goal to improve patient quality of care. Supervision that is needed for the logistic trajectory of the patient on an ad-hoc basis, and is provided directly at the moment of request, significantly enhances the confidence of the resident and improves the quality of patient care. This shows that non-value added waiting time for supervision has a negative impact on quality, which shows the need for supervisors to be available when needed. Kilroy (2006) highlights that supervision aimed at the patient’s trajectory is of vital importance on the way care is provided.

(10)

3. Methodology

To research the relationship between supervision and patient flow in a real-life setting, an embedded case study is held to evaluate factors and effects of changes in the supervision structure. An embedded case study uses multiple sources of evidence with the goal to add depth and richness to the research by integrating qualitative and quantitative data collection methods (Yin, 2003).The sub cases are the surgical and internal medicine (IM) specialty, and the unit of analysis (UoA) is the patient stream of each specialty. An intervention is implemented to improve patient flow by making changes in the way supervision is provided. A before and after study is conducted to determine the effect of changes in supervision structure on patient flow. Data from the pre-intervention study is used to create a new policy in how supervision could lead to an optimal flow. After the implementation of the intervention, the same measurements are taken to determine the outcomes. Observations and interviews are used as qualitative data to interpret the quantitative data.

3.1 Case description

A description of the case is provided, the sub-cases and UoA discussed, and the intervention is described explaining the four changes that are made in the supervision structure.

3.1.1 Martini hospital

This research takes place at the ED in the Martini Hospital in Groningen, located in the north of The Netherlands. This hospital is selected due to its teaching structure where residents perform as first contact for diagnosis and are supervised by specialists. Dynamics like the inflow/outflow patterns, work shift patterns and throughput data of the ED are known by preliminary research by Van Manen (2016) and Gelissen (2016), and so the Martini Hospital provides a basis for this thesis. A description of the hospital and the ED process and lay-out of the Martini is provided in Appendix 3, where is explained how the lay-out of the ED is constructed and used. This describes the setting in which the research takes place.

Patients that enter the ED are assigned to a specialty due to different backgrounds of illnesses and treatment. The four largest specialties with many high acuity patients have their own residents in the ED, being cardiology, lung medicine, internal medicine (IM) and surgery (Gelissen, 2016). Smaller specialties have residents and specialists present within the hospital, but since the smaller number of patients requiring their specific knowledge, their presence is not required 24/7 at the emergency department.

(11)

communicated towards the resident. The notification of arrival provides patient details about the reason of their ED admission. The reason for their announcement is discussed between the specialist and resident, and the patient is placed on the expected arrival list.

In principle, the resident’s perform the treatment of the patients. For high acuity patients in need of immediate care, specialized emergency physicians (EP) are present (one or two, depending on the time of the day and status of the system). EP’s are experienced in trauma situations and possess the knowhow and skillset to deal with acuity. They perform treatment autonomous or together with the corresponding resident of that specialty. The EP’s treat patients if one or more specialties become more crowded, to relieve the system from pressure. EP’s are allowed to treat surgical patients themselves without supervision, if their knowledge and skills are sufficient to cover the (low) complexity of treatment, although the end responsibility still lies with the specialist on supervision duty. This implies the supervisor needs to put his/hers autograph at the treatment policy the next morning when the administrative process is taken care of, although the resident and EP performed the full medical process. EP’s need supervision when treating IM patients and are not allowed to carry the end responsibility of IM patients due to lack of specific knowledge that the specialists do possess. Next to their curing and caring responsibilities, EP’s have the task to coordinate the process with the nurse coordinator. Hierarchically speaking, they are in charge of the logistics of the ED. When flow is stagnating, the EP’s have the duty and the responsibility to coordinate this with the supervisors to achieve improvement in flow (NVSHA, 2013). It is the EP’s end responsibility to keep the flow of the ED as optimal as possible by consulting with all supervisors and residents.

The four largest specialties have their own nursing ward in the hospital. When patients are being hospitalized to one of the nursing wards, the transport is conducted by special ED transport nurses, called ED ward assistants. These nurses provide the transport from and towards the ED. However, when the state of the patient is life-threatening, nursing ward staff has to arrange the transport due to lack of expertise of the ED ward assistant with these patients. In practice, nursing ward staff can transport the patients when there is a lack of availability of staff at the ED due to crowding. A patient can be hospitalized if four tasks have been completed:

- The ‘transfer’ by the ED nurse and the nursing ward nurse, meaning patient information is shared between the ED and nursing ward and responsibility of the patient is transferred. - The policy of CPR is discussed by the resident with the patient/relatives and recorded in the

patient file.

- The medication is submitted in the patient file and approved by the pharmacist nurse. - The medical policy is documented in the patient file by the resident and supervised by the

(12)

3.1.2 Sub-cases

The IM and the surgical specialty are focused on in this study, being the sub cases within the ED of the Martini. The UoA is the patient stream of each specialty. In this thesis, only the day shifts are covered, meaning this will be from the start of the day (9:00) until the day shift switches to the evening shift. Preliminary research of Van Manen (2016) and Gelissen (2016) showed that the amount of patients (WIP) builds up throughout the day, with higher output rates in the evening. By improving the output during the day, this has beneficial effects on the WIP of both day and evening shifts, which is the reason that this research is focused on the day shift.

The surgical specialty can be seen as the department with a high volume of patients combined with a short LoS. IM is the one with long throughput times and a lower volume of patients. The variations between these departments are known and differ substantially, as described in table 1 which is based on research of Gelissen (2016). An elaboration on the description and differences between the specialties is provided in Appendix 4.

Surgical Internal Medicine

Patient percentage of ED 40% 14%

Average LoS (minutes) 141 168

Autonomy of residents High Low

Complexity of treatment Low ( small percentage high) High ( small percentage low)

Mandatory Supervision Only for a small percentage of patients Always

EP end responsibility Yes No

Residents in shift ED 2 1

Table 1: characteristics of specialties.

The specific characteristics of both specialties influence the way supervision is provided. Where for IM residents it is mandatory to ask for supervision for every patient within the process, this only accounts for a small percentage of the patients at the surgical specialty. This is based on the complexity of the patient’s trajectory, which is clarified in Appendix 4. The rest of the patients are treated autonomously by the surgical residents, varying on their experience.

(13)

3.1.3 Supervision Intervention

The intervention comprehends four changes regarding supervision which are applied with the goal to improve patient flow:

- Reduction of waiting time for supervision by establishing a timeframe in which supervision has to be provided.

- Removal of saving up patients (batching) by residents to negotiate together with the supervisor. - Focus on the outflow process.

- Appliance of a Quick Look (QL) Supervision moment.

All these changes are related to each other and all determine how the supervision structure is constructed. These are elaborated down below.

Quick Look and outflow focus

In the pre-intervention process, supervision moments take place whenever the resident requires it, throughout the whole process. After the patient enters the ED, the resident performs the anamnesis, in which the current state of the patient is assessed and its clinical history is observed. After this stage, the resident performs its diagnosis and treatment if necessary. The policy contains the complete diagnosis and treatment plan. Mandatory supervision takes place before or after the policy, depending on the patient’s trajectory, as shown by the blue arrow in figure 5. The end responsibility of the patient lies with the attending supervisor, and so the supervisor needs to approve the policy of the resident. Regular supervision is provided throughout the process whenever the residents require it.

Waiting Room

Enter ED Anamnesis Policy Leave

Figure 5: Pre-intervention process in ED Martini

(14)

Analyzing and discussing the outflow time of hospitalized patients with the management and doctors, showed that a majority of time was waiting time for a bed to be ready for the patient in the nursing ward. In other words, the inpatient departments are not prepared for patients that are arriving on short notice. The nursing wards require more time to prepare for the variability in arrival coming from the ED.

To manage all the previous stated factors, a “Quick Look” stage (QL) is introduced. The residents have to answer three questions after the first time they have seen the patient and discuss this at the earliest possibility with their supervisor:

Does this patient need to be hospitalized?

- When the residents (and supervisor) are certain that the patient needs to be hospitalized, the nursing ward has to be contacted as soon as possible after the quick look, providing them with a larger timeframe to arrange a bed for the patient.

Does this patient need a consult from another specialty?

- If residents from other specialties have to be consulted for their specific expertise, they have to be contacted right away when possible.

What tests (liquid/radiography) does this patient need?

- All the tests that patients would certainly need have to be applied for as soon as possible after the quick look.

The above decision forming has to provide a logistic view at the start of the process for the residents and provide a clinical pathway throughout the ED and hospital for every patient. Structural supervision has to be provided after the QL, as visualized in figure 6. This logistic responsibility of the process is not supported by protocols before, and is a view that especially the inexperienced residents do not fully possess (as discussed and supported by residents and specialists). With this quick look stage, there will be a focus on more efficient logistic care, to implement resources at an early stage as possible. When residents and specialists know at the QL stage that a patient will be hospitalized, they can already start the process of arranging a bed for this patient. This also accounts for consultation from other specialties and time-consuming radiography tests.

Waiting Room

Enter ED Quick Look

Anamnesis Policy Leave

(15)

15 Minute timeframe & Batching

In line with the research of Gelissen (2016), waiting time for supervision is found to be existent in the pre-intervention measurements. With outliers of 50 minutes of waiting, this is found to be unacceptable and needs limitation boundaries for the waiting time to exceed. The management set a 15 minute limit for supervision, meaning specialists are required to provide supervision within a timeframe of 15 minutes after the resident demands it. In the intervention protocol, specialists are instructed that on their supervision duty, they have to be present in the ED even when they are not requested, to ensure the logistic responsibility of the patients. If supervisors have a free moment in their daily activities, they are expected to come to the ED to check up on the residents.

(16)

3.2 Data collection

A before and after study is conducted. The intervention is formed based on pre-intervention measurements. Data from these measurements is used to create new policies in how supervision could lead to a more optimal flow. Post intervention measurements depict the outcomes of the research.

3.2.1 Dataset Throughput times

Pre- and post-intervention throughput times of patients’ processes were collected in the period of 29-8 till 8-9 and 5-12 until 16-12. Disrupting events and special occasions were observed and depicted in the data collection. Measurements lead to patient data of 278 patients which gives insights in their patient journey throughout the ED. This dataset is reduced to 118 patients for which the data was complete. Observatory patients are left out of the dataset since their LoS does not contribute to higher quality of care e.g. patients that need monitoring on their state of health over a longer period of time. Patients who require consultations from other specialties are left out, due to no commitment to the supervision intervention. Data is collected on which event happened at a certain time in the process. The moment that an event takes place is noted down for each individual patient. This is done from the central post of the emergency department, in front of the main screen providing an overview of allocation patient data. Residents provide information about their activities and updates regarding the patient’s state, which ensure for an in-depth understanding of the medical and logistic trajectory. This is conducted for one week for both specialties pre-and post-intervention. The full list of events that are measured can be found in Appendix 1. Data of the first measurements is used to shape the intervention. The systematic choices based on the pre-intervention measurement can be found in the description of the supervision intervention in section 3.1.2. The quantitative data of the surgical specialty did not prove to be reliable enough for this research. This is clarified in the limitations section.

3.2.2 Daily evaluation checklist

To evaluate changes in the structure of the supervision moments, checklists are provided to monitor and identify factors concerning the intervention. Residents note down whenever an event happens during the day that is not part of the protocol, e.g. supervisors not providing them with supervision within the set timeframe. The reason why this happens will be described and evaluated at the daily end meeting of the day shift of the specialty. The checklist and outcomes of the data can be found in Appendix 2.

3.2.3 Observations

(17)

physicians, residents and nursing staff. This was done throughout the ED and hospital, following professionals where needed, involving the central post, treatment rooms, observatory and coffee room. At specific events (high complexity trajectory or crowded situations) the residents provided explanations about strategic choices that were made, to clarify the process and create an in-depth understanding of specific situations. Staff members were interviewed directly or after certain events when they were available for explanation. This was noted down computing a file of field notes. Although their opinions are based on experiences, these are thoughts of professionals with years of experience combined in emergency care, which makes them an authority that ensures the impact of this intervention.

3.2.4 System Throughput times

(18)

3.2.5 Validity and Triangulation

(19)

4. Findings

The multiple sources of data are combined to show the influence by the changes in the supervision structure. These factors are mentioned below and clarified by the combined data. The quantitative and qualitative data of IM is used, where observations and interviews are used to describe the intervention outcomes for the surgeons.

4.1 Early assessment

After the introduction of the new structure, the supervisor had to be contacted after the quick look stage. Depicted below in table 2 is the time it took the residents to contact their supervisor for the first time. After the intervention, a minimum of 0 minutes is measured as it was before the intervention. However, a maximum value of 29 minutes was measured as maximum value. Breaking the 15 minute timeframe only occurred once, due to the arrival of multiple patients at the same time. The second highest value was 12 minutes, showing the immediate use of supervision after the QL at the IM specialty.

N= Minimum Maximum Average Median

Pre-Intervention 18 0 104 23 9,5

Post-intervention 34 0 29 2 2

Table 2: Time in minutes between the end of the first contact moment with patient and first supervision moment In the post-intervention period, supervision is not always provided after the QL. For some patients, supervisors and residents discuss the clinical pathway of patients before their arrival in the hospital at the notification of arrival, and supervision is only provided if their policy has to be changed by new insights at the QL. An example of when supervision was not required at the IM is with patients with a deep venous thrombosis leg (DVT). The trajectory of this diagnosis is low in complexity, and if after the QL the symptoms matched the description, the necessary actions were conducted without contacting the supervisor, since this was all agreed upon at the notification of arrival. This policy holds for both surgical and IM patients.

(20)

Residents, varying from inexperienced to senior, and specialists expected that the LoS has improved after the intervention. They concluded that less time was wasted on non-value added activities. One senior resident with three years of experience in the Martini, who was on supervision duty due to the last year of her specialization, stated:

“The quick look saves up a lot of time compared to how it used to be. It reduces the waiting time for the test results, especially for CT and echography, and the arranging of inpatient beds” (Senior Resident IM)

Especially the requests for a CT-scan and echography were conducted as soon as possible, since these scans require a substantial amount of process time. The faster request of tests saved valuable time in the opinion of the staff, next to the availability of beds whenever the patient was ready to leave the ED.

4.2 Hospitalization Outflow Process

Multiple residents acknowledged that because of the QL, their vision on an emergency situation changed towards a more logistic process. Throughout the process, more logistic awareness was created. On the first day of the new protocol, one of the residents stated:

“I notice I am more aware of the outflow process, so I managed the inpatient process sooner than before. When the patients were ready to leave the ED, they were transported right away. Before the intervention, we had to wait for the bed to be ready at the inpatient ward.” (Resident IM)

If the data is compared for both measurement weeks at the IM, outflow times of ten patients were recorded, which shows the following outflow times (table 3):

N= Minimum Maximum Average Median

Pre-Intervention 10 13 75 40,9 40

Post-intervention 10 3 40 21,3 24,3

Table 3: Outflow time in minutes

(21)

clarified in section 3.1.1. On two of these moments the treatment rooms had to be occupied with new patients, meaning the outflow time had negative consequences for other patients that could not receive treatment due to lack of capacity. The occasion that it took 50 minutes for the patient to leave the ED, there were no patients in the waiting room and other treatment rooms were free. Every action that was necessary from the internal ED staff was done on time, meaning the bottlenecks of the outflow system that add non-value added time, lie outside of the emergency department. In this occasion it was the availability of the nursing ward staff for transferring the patient information.

4.3 Waiting time for Supervision

The time IM residents had to wait to receive supervision is depicted in table 4, starting at the moment when residents took their phone and asked for the specialist to talk about a patient, or when the supervisor was present at the ED and they asked for supervision face-to-face.

N= SV > 15 SV > 30 Average Median

Pre-Intervention 76 8 4 4:39 0:00

Post-intervention 100 2 0 1:03 0:00

Table 4: waiting time for supervision in minutes and number of times waiting time went over 15 and 30 minutes On average, the waiting time reduced with 3 minutes, with the new average being a minute of waiting. This average is lower after the intervention. Occurrence of extensive waiting times were less existent (extreme outliers), with no occurrences of waiting times for more than half an hour. Based on the daily checklists, the amount of waiting time for supervision exceeding the fifteen minute timeframe only happened incidentally in the two month intervention period, showing the significance in reduction of extensive waiting times.

Perceived waiting time became lower, due to their attendance at the ED. This holds for both the surgical department as well for IM:

(22)

Total Batched Percentage

Pre-Intervention 76 42 55,3%

Post-intervention 100 11 11,0%

Table 5: Number and percentage of batched patients

Another factor that reduced non-value added waiting time is a decrease in the batching of patients that was prohibited. Post-intervention, 11% of the IM patients are batched. This is a reduction compared to before, as is shown in table 5. Pre-intervention, residents saved up patients on purpose, for what they believed was for ‘efficiency’ reasons. After the intervention, the batching was done either involuntarily due to same time arrival of patient, or voluntarily and aware of the negative influence of batching process. On a few occasions, batching was deemed to be necessary to improve the quality of care. In multiple cases, supervision was not needed directly within the next hour, and the agreement was made with the supervisor that supervision would be provided when another patient had to be discussed. This implied no non-value added time occurred because of the waiting, since it only improved the efficiency and effectiveness of the specialist’s time.

(23)

4.4 Patient Length of Stay

The IT system used at the Martini (HIX) logs throughput times of every patient entering and leaving the ED. To determine the effects of the intervention on the LoS of the patients, data was retrieved from the system in the pre-intervention as well as the post-intervention period.

All patients IM Surgical Cardiology Lung

2015 188,5 142,4 141,9 156,9 2016 195,2 152,1 154 169,5 Difference 6,7 9,7 12,1 12,5 Increase 3,6% 6,8% 8,5% 8,0% Hospitalized 2015 204,3 184 144,2 165,7 2016 207,7 194,8 155,6 176,7 Difference 3,4 10,8 11,4 11 Increase 1,7% 5,9% 7,9% 6,6%

Table 6: LoS 1 Nov – 16 Dec in minutes

Although the total average LoS became longer for every specialty, there can be seen in table 6 that this increase in time is lower for the IM (3,6%) compared to the rest of the specialties. For patients that are hospitalized, the increase is 1,7 % for the internal department, and the surgical department has a relatively smaller increase (5,9%) compared to the rest of the ED departments. Table 6 shows the averages, however these averages do not fully account for the behavior of throughput times, due to the major influence outliers have on averages. Frequency distribution of LoS provides a clearer image of how the throughput times differ between the pre-and post-intervention periods.

Figure 3: Cumulative frequency distribution of LoS of all specialties in minutes of 1 Nov – 15 Dec 0% 20% 40% 60% 80% 100% 0 50 100 150 200 250 300 350 400 Per ce n tage o f p atien ts

Patient LoS in minutes

LoS All patients

(24)

Figure 3 shows the cumulative frequency distribution of LoS for both time periods. With N being 2500+ for all specialties, results show that the behavior of the LoS of 2015 behaves the same as of 2016. The total LoS of all ED patients become longer, as for every quartile the values became larger. Table 7 depicts the quartile distribution numbers for all departments combined and all the departments individually. Looking at the LoS for all patients (hospitalized and sent home), a smaller increase in the 25% and median quartile is present IM compared to the other departments.

All specialties N = Minimum 25% Median 75% Maximum

2015 2352 14 87 125 175 588 2016 2406 14 94 132 184 606 Increase 54 8% 6% 5% IM 2015 310 34 132,25 178 227,5 484 2016 311 22 138,25 182 241 538 Increase 1 5% 2% 6% Surgical 2015 990 14 85 127 180 588 2016 1113 16 95 134 194 606 Increase 123 12% 6% 8% Cardiology 2015 345 23 101 131 171 376 2016 353 27 112 144 184 420 Increase 8 11% 10% 8% Lung 2015 144 48 118 152,5 192,5 376 2016 146 40 125,25 166 202 419 Increase 2 6% 9% 5%

Table 7: Distribution of patient LoS in minutes of 1 Nov - 15 Dec

(25)

All specialties N = Minimum 25% Median 75% Maximum 2015 761 23 103 139 192 581 2016 796 19 107 152 205 534 Increase 5% 4% 9% 7% IM 2015 156 34 146 191 255,25 484 2016 157 22 151 185 249 538 Increase 1% 3% -3% -2% Surgical 2015 155 48 130,5 169 234 581 2016 158 22 120 181 263 534 Increase 2% -8% 7% 12% Cardiology 2015 154 23 103 137 174,75 356 2016 154 27 112 152 190 307 Increase 0% 9% 11% 9% Lung 2015 85 48 125 157 200 376 2016 85 40 133 174 207 366 Increase 0% 6% 11% 3%

Table 8: Distribution of hospitalized patient LoS in minutes of 1 Nov -15 Dec

To assure that this decrease is due to the intervention, numbers of hospitalized patients clarify if there is an increase in patients and LoS on the nursing wards. Hypothetically, it could be that outflow possibilities to nursing wards were more optimal for IM compared to other specialties and that this is the reason for better outflow performance. In 2016, more patients were hospitalized. Every specialty had an increase, except the Lung nursing department (table 9).

Specialty 2015 2016 Difference Increase

IM 340 361 21 6,2%

Surgical 424 453 29 6,8%

Cardiology 403 443 40 9,9%

Lung 190 189 -1 -0,5%

All 1357 1446 89 6,6%

Table 9: Inpatient admission numbers of 1 Nov-15 Dec

(26)

patients. Figure 4 shows the cumulative frequency distribution of daily amount of hospitalized patients entering the nursing wards. This is for all specialties combined, and shows more daily hospitalizations for 2016 compared to 2015.

Figure 4: Cumulative frequency distribution of patient inflow

4.5 Autonomy and personal characteristics

This thesis identifies factors which are influenced by the supervision structure, however this research showed that the structure itself is also influenced by multiple factors, being the personality and the characteristics of the resident and of the specialist on supervision duty. Their view on a how supervision is dealt with determines how the residents do their job, as discussed in the theoretical section by Loganbill et al.(1982). Multiple IM supervisors who were engaged in the supervision intervention showed great commitment towards the success of the project. With the notification of arrival of a patient they would discuss the logistic steps circumstantially and depict precisely which steps had to be taken for an optimal flow as possible.

While this provides clarity for the patient’s trajectory, it does however raise a question about decision making in the ED. As one supervisor discussed:

“With discussing what steps the residents need to take after the QL, I feel like I am instructing what to do instead of teaching them and letting them figure out the process itself. It downgrades the autonomy of the resident. The hospital is an education platform.” (Supervisor IM)

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 Cumu lat iv e p erce n ta ge o f p at ie n ts h o sp ita lize d at n u rs in g ward s

Amount of patients hospitalized on one day

Inflow inpatient wards

(27)

With this statement, the supervisor implied that when you keep instructing the residents what to do, their training becomes less valuable. There were residents agreeing with this, as one surgeon said:

“I understand the value of the quick look and the “speed” it brings, however I like to work on the

patient myself and when I feel like I need to contact my supervisor, I want to do it. I would rather make choices myself and treat the patient on my own, although it might take longer.” (Resident Surgical)

Although these are beliefs and feelings, it affects the commitment of professionals showing towards the new protocols about supervision. By its multicomponent and subjective nature of definition, it is not strange that professionals have different opinions about how optimal supervision is constructed.

Not everyone agreed with the view that autonomy was changed:

“I notice that with the addition of the quick look and the new supervision structure there is an addition of more overview of trajectory. I don’t feel that my autonomy becomes less, in my opinion that depends on the supervisor and resident themselves. If you discuss the patient together and make choices together, it is beneficial for everyone.” (Resident IM)

Some professionals deemed efficiency and optimal throughput of the ED to be more important than letting the residents works autonomous on patients:

“Sometimes it is better to work as efficient as possible, when the ED is overcrowded, neglecting optimal tutoring of the residents.” (Supervisor IM)

(28)

5. Discussion

The results of this study are discussed in this section, linking the different findings in order to answer the research question. By combining the multiple sources of data and literature are combined, a framework is constructed to visualize how the supervision structure influences patient flow, as is provided on page 29 (figure 5). The relations between characteristics are visualized by numbers in the model. The descriptions of the relations are provided below.

[1] As discussed by Loganbill et al. (1982), the characteristics of the supervisor and resident determine how the supervision structure is constructed. The measurement weeks and daily checklists confirmed this, due to different behavior in coping with supervision by both residents and specialists. Especially the experience of residents play an important role in this (clarified in Appendix 2). Next to this, the strategic choices that are made by management serve as a guideline how supervision should be provided and how autonomous residents are allowed to work. These two combined, with other factors as the environment and patient characteristics, form the degree of autonomy that a resident has on the ED process of a patient and how supervision should be structured. Section 4.5 covers the view of residents and specialists on autonomy and strategic decisions. Since all these factors are interrelated, this is visualized as the purple box in figure 10.

[2] During the measurement weeks, the experience of residents showed that the autonomous work structure can have beneficial, as well as negative influences on patient flow. Experienced residents are bound by restrictions if waiting for supervision is required for certain handlings, hence reducing flow. More negative consequences appear when an autonomous resident does not know how to cope with certain situations and does not ask for supervision. Beneficial outcomes, such as the autonomy to work on lower complex patients, as with the surgical department, optimize patient flow by removing unnecessary waiting time for supervision.

(29)

[4] As provided and confirmed by the residents, the new protocols from the intervention helped them with creating more logistic awareness and responsibility. This is also influenced by the experience of the resident; the more experienced the residents are, the more logistic responsibility they possess, as came forward in the measurements and daily checklists.

[5] Due to discussing the trajectory of the patient with their supervisor early on in the process (being the notification of arrival or after the quick look; early assessment) the residents were more aware of the trajectory of the patients. This increase of logistic responsibility helped the residents obtain a (more) helicopter-view approach towards the process, creating more feeling for patient flow in the system. The early assessment and outflow coordination is determined by the supervision structure. The logistic responsibility of the residents enhances the use of this. This does not account for every patient, since for certain complex symptoms the residents nor specialists were still not able to provide a clear trajectory after the QL. This applies for IM due to the high complexity of patients’ illnesses, but also for surgery on a smaller scale.

[6] The effect of the early assessment and improved outflow coordination is a reduction of non-value added time, as described in section 4.1, 4.2 and 4.4. In the pre intervention week, the majority of hospitalization outflow time was due to waiting for the nursing ward to free up resources for the admission of the patient, as in jargon would be said: “free up a bed”. In the post-intervention week, all the beds that needed to be arranged on the nursing ward were arranged in the beginning of the process. As a consequence, the nursing ward itself was prepared for patients to be transferred when the ED was ready to admit the patient, meaning the outflow times in this week were not due to waiting for the ward to free up a bed as it was before. The pre-intervention measurements showed that out of the ten hospitalized patients, four of them did not managed to leave the ED as soon as their ED process was finished, implying that other bottlenecks are present in the outflow process.

[7] The logistic responsibility is a key factor determining the flow of the ED. An increase in logistic responsibility results in better performing flow. Residents acknowledged that due to the intervention protocols, their logistic responsibility became larger. It helped them identify and reduce non-value added time during early stages of the process, as well as the outflow process of patients that are hospitalized.

(30)

The consequence of optimization of patient flow is the reduction of patient LoS. The quantitative data of improvements in LoS for IM can be related to the intervention. Table 9 and figure 4 depict that the overall activity when it comes to hospitalizing patients becomes higher, at the surgical, IM and cardiology specialties. Due to higher inflow (TH), more patients enter the system. Applying Little’s Law (Hopp, 2011), WIP = TH X CT, the amount of patients (WIP) becomes larger. In theory, the LoS (CT) of inpatients does not change. However, discussing the data of Hotflo with the integral capacity manager, it showed that when nursing units become more crowded, inpatient LoS becomes lower. In this research, data of inpatient LoS is unknown, so there cannot be guaranteed to a full extent that the WIP in 2016 is larger compared to 2015. It could be possible that outflow for hospitalized IM patients improved since fewer patients were present at the IM nursing ward, so more capacity was available for hospitalization compared to other specialty nursing wards. To completely verify that the improved ED LoS of IM is improved all due to the intervention, quantitative data of inpatient LoS needs to be available. However, both nursing ward staff and ED staff agreed that due to the intervention, the hospitalization process is improved for the IM specialty, resulting in a more optimal LoS.

Supervisor & Resident Characteristics Supervision Structure Patient Flow Logistic Responsibility Autonomy of Resident Clear Pathology: Early Assessment/ Outflow Coordination Reduction of Non-value added waiting time Strategic choices defined by Management 1 5 4

+

3 2 8 7 7 6

Figure 5: Visual model relating the supervision structure to patient flow

(31)

6. Conclusion

The aim of this thesis is to provide an answer to the following research question:

How does the supervision structure influence the flow of an emergency department?

The supervision structure in an ED can influence flow by reducing non-value added waiting time. By focusing on a fixed timeframe in which supervision has to be provided, extensive waiting times decreases and the batching of patients ceases to a minimum, and so reduce the average and median values for supervision waiting time. The supervision structure contributes on laying the emphasis on logistic decision making, which is the key to avoiding non-value added waiting time. If residents obtain a clear view on the trajectory of a patient, which can be provided and assured by adequate supervision, there can be assessed at an early stage which resources patients require in their process. Radiography tests and consults from other specialties can be applied for in an early stage, and informing the nursing ward for patients that require to be hospitalized can be arranged as soon as possible. This leaves more time for other departments (nursing ward, radiography) to arrange capacity that is needed for the patient. Supervision can enhance the logistic responsibility of residents, and has beneficial as well as negative influences on the autonomy of the residents. Supervision can be used in situations of crowding, by reducing the educational side of the ED and using the experience of specialists on focusing on efficiency and productivity when this is required.

In healthcare, changes need to comply with standards of the triple aim: improving the patient experience, improving the health of a population, and reducing costs per capita (Berwick, Nolan & Whittington, 2008). This research shows the patient experience can be enhanced by reducing non-value added waiting time and improving patient LoS. The ED patient flow can be optimized, showing beneficial outcomes for other ED patients as population benefits. No extra staff was needed in providing supervision, implying no additional staffing costs are inflicted.

(32)

7. Limitations & Further research

To fully verify the reduction in LoS for IM inpatients, LoS data of the nursing wards need to be available, as is clarified in the discussion section of this thesis. This is a limitation in this research due to lack of time and resources. Future research could verify the intervention improvements by linking the ED LoS to nursing ward LoS, and conduct the same research over a longer period of time with the same intervention protocols as in the post-intervention period.

During crowded moments, the author was not able to register all the actions of the surgical medical staff during the measurement weeks, nor were the residents able to fill in the daily checklists for evaluation at certain days. As a consequence, the quantitative data of the surgical specialty was incomplete. In peak moments, four to five residents were working inside the ED and multiple specialists provided supervision together with the EP’s. The physical distance between their workspaces made it impossible for the author to track all of the resident’s and specialist’s actions. There was a lack of commitment towards the intervention protocols from certain surgical staff members, not following the protocols of the intervention. This makes it hard to relate the outcomes of quantitative results with the intervention, which is the reason the quantitative data of the surgical department is left out of the findings. Future research requires more manpower than one observer to fully cover the ED, and address the importance of the supervision structure at the surgical specialty when it comes to patient flow. The improvements at IM can serve as a guideline for implementing a strategy that can help specialties such as surgery with improving patient flow.

(33)

8. Managerial Recommendations

One of the primary outcomes for optimizing LoS throughout the system is the logistical responsibility of the staff working in the ED. More experienced staff has a better insight in how the logistics of the entire department behaves. Especially the EP’s have a good insight in this, as is described in their function (NVSHA, 2015). As a recommendation, the surgical residents need to ask for supervision after their QL at the EP’s. Research addresses the importance of this as improvements in diagnostics and treatment are noticed when residents call for supervision directed at emergency physicians (Callander, 2009). As this research points out, the supervision structure at the surgical department is not optimal as it is now, with waiting times for supervision still present due to the dynamic structure of specialists on supervision duty. A standard protocol for QL supervision directed at the EP’s should be a better solution for this department, as the physique presence of EP’s is optimal for supervision provision compared the absence of specialists. Regular check-up moments by the EP’s at crowded situations could play an important role in this, maintaining at least one person who keeps a helicopter-view of the total state of the ED process. This could be every hour or two hours, depending on the state of crowding or the potential of the ED to become overcrowded. This requires ad-hoc solutions for crowded situations and plays an important role for short term problems. To create a solution on the long term, suggestions are to create more logistic knowledge and responsibility for the residents and potentially the entire ED staff. By providing the residents with logistics training, they raise more awareness for crowding. Teaching the staff about supply chain behavior of the ED, helps them create this responsibility and awareness. Where most medical professionals obtain a high degree of clinical responsibility, more logistic responsibility should be the effect of these trainings.

This research shows the relationship between protocols describing how the patient process should be addressed, and the following of this by residents through supervision. This was supported by the research of Kingsbury and Allssop (1994). Protocols are changed into structural early assessment as was done for the IM. If the adequate supervision for this is provided, this would imply a larger focus on patient logistics. This research showed the beneficial relationship between logistic responsibility and patient LoS, and so should be implemented at the cardiology and lung department.

Management needs to clarify what function the ED carries regarding the teaching. The doubt of multiple professionals about autonomy implies that management of the ED and the hospital needs to evaluate the impact of the change in policy and protocols about optimal flow on the autonomy of residents and needs to find a funded answer for the question: When is the ED a teaching institute (

(34)

9. References

American College of Emergency Physicians. (2006). Crowding. Annals of Emergency Medicine, 47:585.

Asplin, B. R., Magid, D. J., Rhodes, K. V., Solberg, L. I., Lurie, N., & Camargo, C. A. (2003). A conceptual model of emergency department crowding. Annals of emergency medicine, 42(2), 173-180.

Berwick, D. M., Nolan, T. W., & Whittington, J. (2008). The triple aim: care, health, and cost. Health

affairs, 27(3), 759-769.

Carter, E.J., Pouch, S.M. & Larson, E.L., 2014. The relationship between emergency department crowding and patient outcomes: A systematic review. Journal of Nursing Scholarship, 46(2), 106–115.

Callander, M. E. (2009). Literature Review of emergency department Staffing Redesign Frameworks (Doctoral dissertation, University of Sydney).

DeLaney, M., Zimmerman, K. D., Strout, T. D., & Fix, M. L. (2013). The effect of medical students and residents on measures of efficiency and timeliness in an academic medical center emergency department. Academic Medicine, 88(11), 1723-1731.

Derlet, R. W., & Richards, J. R. (2000). Overcrowding in the nation’s emergency departments: complex causes and disturbing effects. Annals of emergency medicine, 35(1), 63-68.

Driscoll, J. (2007). Practicing clinical supervision: a reflective approach for healthcare professionals. Elsevier Health Sciences.

Drupsteen, J. (2013). Treating planning flaws in patient flows (Doctoral dissertation, University of Groningen).

Eva, K. W., & Regehr, G. (2005). Self-assessment in the health professions: a reformulation and research agenda. Academic Medicine, 80(10), S46-S54.

Hoot, N. R., & Aronsky, D. (2008). Systematic review of emergency department crowding: causes, effects, and solutions. Annals of emergency medicine, 52(2), 126-136.

Hopp, W. J. (2011). Supply chain science. Waveland Press

Ganguly, S., Lawrence, S., & Prather, M. (2014). Emergency Department Staff Planning to Improve Patient Care and Reduce Costs. Decision Sciences, 45(1), 115-145.

(35)

Jha, R. K., Sahay, B. S., & Charan, P. (2016) Healthcare operations management: a structured literature review. DECISION, 1-21.

Joosten, T., Bongers, I., & Janssen, R. (2009). Application of lean thinking to health care: issues and observations. International Journal for Quality in Health Care, 21(5), 341-347.

Kilminster, S. M., & Jolly, B. C. (2000). Effective supervision in clinical practice settings: a literature review. Medical education, 34(10), 827-840.

Kilminster, S., Cottrell, D., Grant, J., & Jolly, B. C. (2007). AMEE Guide No. 27: Effective educational and clinical supervision. Medical teacher, 29(1), 2-19.

Kilroy, D. A. (2006). Clinical supervision in the emergency department: a critical incident study.

Emergency medicine journal, 23(2), 105-108.

Kingsbury, S., & Allsopp, M. (1994). Direct consultant supervision of higher trainees in child and adolescent psychiatry. The Psychiatrist, 18(4), 225-229.

Kinsman, L., Rotter, T., James, E., Snow, P., & Willis, J. (2010). What is a clinical pathway? Development of a definition to inform the debate. BMC medicine, 8(1), 1.

Ladany, N., Mori, Y., & Mehr, K. E. (2013). Effective and ineffective supervision. The Counseling Psychologist, 41(1), 28-47.

Loganbill, C., Hardy, E., & Delworth, U. (1982). Supervision: A conceptual model. The Counseling

Psychologist.

Moss, J. E., Flower, C. L., Houghton, L. M., Moss, D. L., Nielsen, D. A., & Taylor, D. M. (2002). A multidisciplinary Care Coordination Team improves emergency department discharge planning practice. Medical Journal of Australia, 177(8), 435-439.

Nugus, P., Holdgate, A., Fry, M., Forero, R., McCarthy, S., & Braithwaite, J. (2011). Work pressure and patient flow management in the emergency department: findings from an ethnographic study.

Academic Emergency Medicine, 18(10), 1045-1052.

Nederlandse Vereniging van Spoedeisende Hulp Artsen (2013) Standpunten NVSHA t.a.v. opleiding,

organisatie acute zorg en beroepsbelangen

http://www.nvsha.nl/images/nvsha/standpunten/20130808_standpunten_NVSHA.pdf

Nederlandse Vereniging van Spoedeisende Hulp Artsen (2015) Richtlijn Workforce SEH

(36)

Pines, J. M., Garson, C., Baxt, W. G., Rhodes, K. V., Shofer, F. S., & Hollander, J. E. (2007). ED crowding is associated with variable perceptions of care compromise. Academic Emergency

Medicine, 14(12), 1176-1181.

Regionale Overleg Acute Zorg (2016) Brandbrief ‘Nationale spoedzorg - de rek is er uit’

http://www.omroepflevoland.nl/SiteFiles/Doc/brandbrief_regionale_spoedzorg_-_de_rek_is_er_uit_2016-05-19_F5F6C9FFA1430EA2C1257FBC00329FBF.pdf

Salvi, F., Morichi, V., Grilli, A., Giorgi, R., De Tommaso, G., & Dessi-Fulgheri, P. (2007). The elderly in the emergency department: a critical review of problems and solutions. Internal and Emergency

Medicine, 2(4), 292-301.

Schippers, E., (2016) Schippers onderzoekt weigering spoedeisende hulp drukte.

https://www.rijksoverheid.nl/binaries/rijksoverheid/documenten/kamerstukken/2016/06/06/kamer brief-over-bericht-brandbrief-regionale-spoedzorg-de-rek-is-eruit.pdf

Sloan, G., & Watson, H. (2002). Clinical supervision models for nursing: structure, research and limitations. Nursing Standard, 17(4), 41-46.

Snowdon, D. A., Hau, R., Leggat, S. G., & Taylor, N. F. (2016). Does clinical supervision of health professionals improve patient safety? A systematic review and meta-analysis. International Journal

for Quality in Health Care.

Tang, C., Chen, Y., & Lee, S. (2015). Non-clinical work counts: facilitating patient outflow in an emergency department. Behaviour & Information Technology, 34(6), 585-597.

Thijssen, W. A., van Miero, E., Willekens, M., Rebel, J., Sandel, M. H., Giesen, P., & Wensing, M. (2015). Complaints and diagnoses of emergency department patients in the Netherlands: a comparative study of integrated primary and emergency care. PloS one, 10(7), e0129739.

Thomas, J. W., Guire, K. E., & Horvat, G. G. (1997). Is patient length of stay related to quality of care?

Journal of Healthcare Management, 42(4), 489.

Van den Brink, R. & Van der Parre H. (2011) Helft spoedeisende hulp kan dicht http://nos.nl/artikel/235398-helft-spoedeisende-hulp-kan-dicht.html

Van der Linden, C., Reijnen, R., Derlet, R. W., Lindeboom, R., van der Linden, N., Lucas, C., & Richards, J. R. (2013). Emergency department crowding in The Netherlands: managers’ experiences.

(37)

Van Manen, V. (2016) Emergency care waiting times; The impact of dynamics and room utilization. University of Groningen

Voss, C., Tsikriktsis, N., & Frohlich, M. (2002). Case research in operations management. International

journal of operations & production management, 22(2), 195-219.

Wiler, J. L., Gentle, C., Halfpenny, J. M., Heins, A., Mehrotra, A., Mikhail, M. G., & Fite, D. (2010). Optimizing emergency department front-end operations. Annals of emergency medicine, 55(2), 142-160.

White, A. L., Armstrong, P. A., & Thakore, S. (2010). Impact of senior clinical review on patient disposition from the emergency department. Emergency Medicine Journal, 27(4), 262-265.

(38)

Appendix 1. Dataset of throughput times

The time of the following events were noted down in Microsoft Excel to compute a dataset of throughput times (table 10).

Patient Arrival Waiting Room Patient Arrival ED

Start Anamnesis / Quick Look Stop Anamnesis / Quick Look

Hospitalization: yes/no/cannot be assessed Request for Supervision

Start Supervision Stop Supervision

Kind of Supervision: phone/computer/bed Waiting time for Supervision

Duration Supervision Start Outflow Outflow Moment Duration outflow ED LoS Hospitalization

Specialty, Specialist, Resident Illness + Triage Color

Extra Notes

Table 10: Measurement time moments

(39)

The next measurement was the first time that residents had contact with their supervisor and how long they had to wait for it. Also, the sort of supervision was noted, being by phone, joined behind a computer negotiating the patient’s status or together at the patient itself in the treatment room. All the other times residents contacted their specialist for supervision were registered on the same way as the first supervision moment, providing a complete list of all the supervision encounters between supervisor and residents of that week. The time residents had to wait to actually get supervision was depicted, starting at the moment when residents took their phone and asked for the specialist to talk about a patient.

(40)

Appendix 2. Daily evaluation list of Supervision

(41)

(42)

For every IM patient not needing a supervision moment after the quick look, the residents noted down they did not require supervision and the reason why. From all these patients, close to a third of the patients did not need supervision because all the logistic choices that can be made at the notification of arrival were not varying as they were discussed at the announcement of the patient. The sum of the amount of patients requiring supervision after their QL (by supervisor/EP), together with the patients not needing supervision after the QL, are noted down by the residents as shown in figure 7. Both lines are almost equal, showing the high utilization of the quick look and the determination of the residents of the internal medicine administrating their efforts. Where the green line becomes higher than the red line (i.e. more supervision is received than patients were present), this is due to receiving supervision from both the supervisor as well as the emergency physician.

Figure 7: Quick look supervision moments for IM based on daily checklists

From this data, the majority of the patients received a QL within 15 minutes or did not need a quick look supervision moment (as was noted down by the residents). This has to do with multiple variables, of which the most important one is the nature of the patient’s illness and treatment complexity. High complex patients were contacted right away after the QL, where patients with minor complexity and clear trajectory required less supervision. Low complex patients did not change in policy after the quick look compared to the notification of arrival.

0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 N u m b e r o f e n titi e s Number of Day

Quick Look Internal Medicine

Amount of daily patients

Amount of QL Supervision moments

The impact of the supervision structure on Emergency Department flow.