Diagnosis of patient throughput times At the surgical department of a hospital

(1)

Master thesis

Faculty of Economics and Business

MSc Business Administration – Operations & Supply Chains

13-07-2012

Jan Pieter Bos

Student number: 1765000

Koningsschut 17 8309 CC Tollebeek The Netherlands j.p.bos@student.rug.nl

Diagnosis of patient throughput times

(2)

1. Introduction

This thesis is written as part of the Master Business Administration - Operations & Supply Chains at the University of Groningen. I conducted an investigation at a medium-sized hospital with a regional care function. The hospital asked for a diagnosis of delaying factors in the care path of surgery patients, specifically surgery patients who are treated at the radiology department. The hospital management feels that reduction of waiting times is needed, because they are perceived as being too long by different stakeholders. However, the hospital management does not have a clear picture of all processes and patient flows in the hospital. When planning adjustments have to be made, the displacement effects on the less urgent processes are not really clear. Furthermore, the patient throughput times are unclear. My investigation is part of a larger investigation carried out by the hospital to find a solution for the problems described above.

The objective of this research is to help the hospital management with improving the processes that take place in the surgical department. To enable the hospital management to find improvement opportunities, I will make a structured diagnosis of the current performance in the hospital using the available literature. The patient throughput times will be taken as the basis of this diagnosis. This is the time required for a whole series of activities to be carried out from the moment the patient contacts the hospital to the moment the patient is treated. To achieve this objective and to answer the problem described above, the following research question will be answered:

Which factors are the main factors of throughput time delays in the care path of patients in the surgical department of the hospital?

The research derives its practical relevance from its support to managerial decisions in the researched hospital. Its scientific relevance stems from the diagnosis framework applied, which is new in research on patient flows.

(4)

2. Literature

The main purpose of this chapter is to discuss the relevant academic literature for this research. In the first part, a literature review will be done and in de second part a conceptual model will be presented. The literature review is about patient throughput times in hospitals.

2.1. Literature review

Patient throughput times are affected by waiting times. If patients in a hospital have to have elective surgery, they may be confronted with long waiting times. Therefore, waiting times are an important health policy issue. This is, among other things, the case in the Netherlands (Siciliani and Hurst, 2003). Different policies are introduced to solve this problem. For example, one of the policies introduced is that patients have more opportunities with regard to choosing their hospital. This creates more competition between hospitals (Ringard and Hagen, 2011). But reducing waiting times is not easy to achieve because there is a complex relationship between waiting times and activities for individual surgical procedures (Siciliani and Hurst, 2003).

In order to improve waiting times, in the past much research was done in small-sized projects in hospitals by focusing on specific small groups of patients. In such situations improvements can generally be realized by giving the group in question a higher priority. For example, a lot of research has been done in emergency departments (Ekelund et al. (2011), Cassidy-Smith (2007), and Takakuwa et al. (2007)). But there is still a gap in literature about reducing throughput times in general or for a relatively large group of patients. More specifically, for this research I need a model for structured diagnosis of the throughput time performance in hospitals, but this is not available in the literature. The only usable comprehensive framework found is Soepenberg et al. (2012). But this model is developed for the order flows of make-to-order (MTO) companies and not for hospital environments. The authors present a comprehensive framework to diagnose delivery reliability in MTO companies. This model aims to assist in identifying those production planning and control (PPC) decisions that are negatively influencing the delivery performance. The first part of the framework determines the relevant problem areas and consists of four steps:

Step 1: Analyse the distribution of lateness. Step 2: Analyse differences among order subsets. Step 3: Analyse differences over time.

Step 4: Analyse delivery-time-promising process and the realisation process.

The second part of the framework is determining PPC causes within the problem areas found in the first step.

(5)

lateness and the variance of lateness in a company it is important to have data about promised and actual delivery times. This also applies when the diagnosis is done in a hospital. In order to apply the diagnosis model to hospital settings it is important to have set time frames within which a patient should be treated. How I am using the framework in a hospital situation is explained in chapter 3. This research is about making a diagnosis. The factors influencing each other and finally the throughput times in the hospital are not known and have to be found using the diagnosis. But the literature can guide us in searching for these factors. Therefore I use Ling and Benton (2002), which is about hospital capacity management decisions. Hospital services are essentially capacity driven. The capacity consists of resources such as facilities, equipment, and workforce. This influences the services offered, the quality of service, the cost of service, and the satisfaction of customers, employees and payers. This information can be used to build a conceptual model, which is done in the next subsection.

2.2. Conceptual model

I made a conceptual model to provide an overview of the factors influencing each other and, finally, the throughput times in the hospital. These factors are found in the literature and serve as a guide when following the model of Soepenberg (2012). This conceptual model is made using the interviews I had at the hospital and the literature described above, see figure 1. Most factors used in this conceptual model are found in Ling and Benton (2002). The factors “Agreements” and “Planning issues” are were derived from interviews at the hospital.

Figure 1: Conceptual model

Explanation of the conceptual model:

The process is the series of activities to be carried out in the hospital from the moment the patient contacts the hospital to the moment the patient is treated. The process is influenced by the demand

Processes

Agreements

Throughput time

Demand for care

(6)

(7)

3. Method

This section presents an overview of the methods used in this study. This research will focus on the activities of patients treated at the radiology department with as applicant the surgeon. This means that the patients to be investigated within the scope of this research are of the surgical specialty and are all treated at the radiology department (hereafter referred to as surgery patients). This is a conscious framing of this research because the hospital asked for a diagnosis of delaying factors in the care path of surgery patients who are treated at the radiology department. As already explained in the introduction, the hospital management feels that reduction of waiting times is needed, because they are perceived as being too long by different stakeholders. However, the hospital management does not have a clear picture of all processes and patient flows in the hospital. When planning adjustments have to be made, the displacement effects on the less urgent processes are not really clear. Furthermore, the patient throughput times are unclear. My investigation is part of a larger investigation carried out by the hospital to find a solution for the problems described above. To be able to compare the patient journey of the surgery patients to the rest of the radiology patients, I have also investigated the performance of the entire radiology department.

First, the hospital and in particular the radiology department is described. The second subsection describes the data collection. The third section explains how I plan to carry out the diagnosis involved in this research.

3.1. The hospital

This research was done at a hospital with all specialty services available, 339 beds, and extensive outpatient facilities. Quality of care is very important for this hospital. So much attention is given to the best quality of care. The hospital has a large number of quality marks and accreditations.

Radiology is a medical specialty that uses ‘imaging’ studies on the inside of the human body to make a good diagnosis. In order to get an impression of the processes in the department, it is described below. Different techniques are used to make diagnoses: General radiology, Vascular interventional radiology / angiography, Computed tomography (or CT) examination, Mammography, Ultrasound, Magnetic Resonance Imaging (MRI), and Nuclear medicine. The department consists of several sub departments: Angiography, Bucky, CT, Dex, Screening, Echo, Interventions, Mammography, MRI, and Nuclear Medicine. The department has seven radiologists, one department head, three coordinators, one practice trainer, thirty-five Medical Imaging and Radiation Experts (MIRE), two MIRE’s in training (dual), and five medical secretaries. Furthermore, two digital Bucky rooms and one trauma room, two digital screening devices, three Echo devices, a Mammography device, a CT, a MRI, a mobile screening device, a mobile X-ray machine, and a Gamma camera are available.

3.2. Data collection

(8)

computer systems of the hospital. I chose to do the data collection in two phases: the first phase resulting in preliminary analysis which made clear what additional data (the second phase) was needed for the final analysis.

A selection was made of all patients and their operations at the radiology department within the period of 2010 to April 2012. From these data a selection was made from the period of 1 July 2010 to 30 June 2011, which is mostly used in this research because these data contain the most complete patient journeys. The other data are only used for building throughput diagrams. Also, a sub selection was made from the same periods of all patients at the radiology department with as applicant the surgeon. From these patients all operations associated with a surgical Diagnosis-related group (DRG) are selected. This sub selection provided more information, such as DRG numbers, type of care, primary diagnosis codes and treatment codes, and was made to enable a more in-depth research on the activities of patients treated at the radiology department with as applicant the surgeon. Activities of the Bucky department are not included in this research because these operations done using a walk-in system in which almost only capacity is planned.

In addition to data collection as described above, I gathered information during conversations held with the project team of the hospital, consisting of people related to this research. During these conversations we explored the problem in the hospital as it is perceived by the management. Also, interviews with an IT employee were carried out to clarify the availability of relevant quantitative data in the computer systems of the hospital. Furthermore, I interviewed the planning staff about the planning system in the hospital. Once I had some results during the diagnosis process we discussed these results with the project team to get feedback for further investigation.

The website of the hospital was also used for gathering general information about the hospital.

3.3. Data analysis

The data analysis is divided into two phases: a preliminary analysis and the final analysis. The preliminary analysis includes an investigation of patient throughput times, a comparison of different departments in the hospital, and a first investigation into the waiting times in the hospital. In the final analysis the framework of Soepenberg et al. (2012) is applied step by step as described in the paragraphs below.

The framework of Soepenberg et al. (2012) discussed in chapter 2 will be used as a starting point and adapted for hospital environments. The different steps described in the framework will be followed. During these steps some decisions need to be made for adapting the framework to hospital environments. This is done using the available data. When the framework says ‘orders’ this means ‘patient journey’ in hospital environments. How I applied the different steps of the framework is described below. All steps are done for the entire radiology department and specifically for the surgery patients.

Part 1: Determining the relevant problem areas.

(9)

In this step, I investigate duration instead of lateness (as it says in the original framework). I look at the duration of the patient journey in the hospital. Since there are no promised throughput times available I only investigate the realized throughput times. The realized throughput time is the time between the application and the operations. I showed these throughput times to the hospital management to see if there are unexpectedly long durations.

Step 2: Analyse differences among subsets

In step 2 I investigate differences among subsets in the hospital instead of orders in the original framework. These subsets are the different radiology sub departments, different patient groups regarding urgency (only for surgery patients because I don’t have data regarding urgency for all radiology patients), and different specialties. When subsets have high internal variance I continue with investigating within these subsets, which is called variance-oriented diagnosis. If there are mainly differences between the subsets I will continue with investigating those subsets where the duration is long, which is called average-oriented diagnosis.

Step 3: Analyse differences over time

Differences over time will be investigated in this step. This means that I will investigate the throughput times over time to reveal differences between specific time periods. As I already explained in the first step I will not look at lateness but at duration.

Step 4: Analyse patient-journey-time-promising process and/or the patient journey process

This step is done to determine whether the diagnosis should continue in the process of patient-journey-time-promising and/or in the patient journey process itself. This means that as an adjustment to the original framework I would use ‘patient-journey-time-promising’ instead of ‘delivery-time-promising’ and ‘patient journey process’ instead of ‘realization process’. In case of average-oriented diagnosis (see step 2), throughput diagrams are used for gaining insight. In case of variance-oriented diagnosis order progress diagrams are used to make the decision. However, I do not have any information available about promised treatment times. Therefore, it is hard to investigate the patient-journey-time-promising process. So, only the patient journey process will be investigated.

Part 2: Determining PPC causes within problem areas

(10)

4. Results

This chapter presents the results of the diagnosis done, using the framework described in the previous chapter. As explained in chapter 3, this research will focus on the activities of patients treated at the radiology department with as applicant the surgeon. This means that the patients to be investigated within the scope of this research are of the surgical specialty and are all treated at the radiology department. But, to have a good insight into the performance of the radiology department, I first apply the framework to all patients of the radiology department (chapter 4.1). In chapter 4.2 I zoom in on the surgery patients.

4.1. Radiology – all patients

In the next subchapters the results of the diagnosis  using the framework of Soepenberg (2012)  of all radiology patients are presented.

4.1.1. Step 1: Analyse the distribution of duration

In order to see the distribution of the duration I made an overview of the time between the application and the operation itself of all operations at the radiology department. 5929 patients (18,79 percent) were treated the same day as the application was filed, 2584 patients (8,19 percent) were treated one day after the application, and 2076 patients (6,58) were treated two days after the application. The rest of the patients and the number of days they were treated after application are presented in figure 2.

Figure 2: Time between application and operation

The number of days between the application of an operation and the operation itself. The data relate to the number of patients in the period of 1-07-2010 to 30-06-2011.

(11)

This figure combines all kinds of patients and all kinds of operations. However, it is possible that there are differences among subsets.

4.1.2. Step 2: Analyse differences among subsets

In the previous subchapter I showed that a lot of patients are treated very quickly after they file an application. In order to gain more insight into this phenomenon I investigated the different radiology sub departments and the different specialties which use the department of radiology. In this way, I can learn whether there are differences between these groups.

4.1.2.1. Differences between radiology sub departments

In figure 3 the number of patients is shown for each radiology sub department. The Echo department treated more patients than the other departments. Also, CT and MRI treated a lot of patients, OK treated only 37 patients. These differences can be explained by the nature of the treatments done in these departments.

Figure 3: Number of patients per sub department

The number of patients treated per radiology sub department. The data relate to the number of patients in the period of 1-07-2010 to 30-06-2011. The sub departments have Dutch names: Angiografie for Angiography, CT for Computed Tomography, Dex for Dexa (bone density measurement), Doorlichting for Screening, Interventies for Interventions, Mammografie for Mammography, MRI for Magnetic Resonance Imaging, Nucl. Geneesk. for Nuclear Medicine, and OK for Operating Room.

(12)

a treatment in 0, 1 or 2 days for most of their patients. MRI, which is an expensive resource, is not that fast. A lot of patients can go there only after two or more days. Echo, which treated the largest number of patients, is not the fastest department but hospital personnel did not find this notable. The differences between the departments can be explained by the nature of the treatments done there. The conclusion that can be drawn is that MRI is remarkably slow and should be further investigated.

Figure 4: Throughput time per sub department

Throughput time per radiology sub department, displayed in categories of the number of days (see the legend). The data relate to the number of patients in the period of 1-07-2010 to 30-06-2011. The sub departments have Dutch names: Angiografie for Angiography, CT for Computed Tomography, Dex for Dexa (bone density measurement), Doorlichting for Screening, Interventies for Interventions, Mammografie for Mammography, MRI for Magnetic Resonance Imaging, Nucl. Geneesk. for Nuclear Medicine, and OK for Operating-room.

4.1.2.2. Differences between specialties

(13)

Figure 5: Number of applications per specialty

The number of patients treated per specialty. The data relate to the number of applications in the period of 1-07-2010 to 30-06-2011. The specialty codes are: ANA for Anesthesia CAR for Cardiology, CHI for Surgery, DER for Dermatology, GER for Geriatrics, GYN for Gynecology, IC for Intensive Care Unit, INT for Internal Medicine, KIN for Pediatrics, KNO for Otorhinolaryngology, LONG for Long Medicine, MDL for Gastrointestinal and liver, MHK for Oral Surgery, NCH for Neurosurgery, NEU for Neurology, OOG for Ophthalmology, ORT for Orthodontics, PLAS for Plastic Surgery, PSY for Psychiatric, RAD for Radiology, REU for Rheumatology, REV for Rehabilitation, and URO for Urology.

Figure 6: Throughput time per specialty

(14)

4.1.3. Step 3: Analyse differences over time

The next step in this diagnosis is analyzing time dependency. In figure 7 the throughput diagram is shown. I will first explain how to read throughput diagrams in this research. The throughput diagram contains cumulatively all operations (the output curve) with the corresponding applications (the input curve). The curves increase respectively in value by the amount of operations and applications executed at each date on the horizontal axis. The vertical distance between the input curve and the output curve depicts the work-in-progress (WIP), and the horizontal distance relates to the average throughput time when the patients would have gone through the department in a first-come-first-served sequence. In the throughput diagram only averages are shown. For individual patients the throughput time can be different. Noticed that the average throughput time is influenced by outliers, in the sense that patients with extreme long throughput time influence the distance between the curves for long periods. Also, there is a small lack of data because the applications (and their corresponding operations) filed 9 months or more before April 2012 are not included in this throughput diagram. All types of operations are included in this throughput diagram, even follow-up appointments, in order to have the most complete capacity load and because the follow-up appointments could not be excluded using the available information.

The curves run almost parallel upward even when I zoom in per month. The WIP and the throughput time vary over time but not with remarkable quantities (10 percent or less from average). The large number of patients cause a relatively regular arrival pattern. Therefore, there is no large capacity flexibility needed on the aggregated level.

Figure 7: Throughput diagram of applications and operations at Radiology

(15)

4.1.4. Step 4: Analyse the patient journey process

As explained in chapter 3, in this step of the diagnosis I can only decide to investigate the patient journey process. Using throughput times I can make some general conclusions about the performance of the patient journey processes at the radiology department. Therefore, I will continue with this in part two of this diagnosis.

4.1.5. Part 2: Determining PPC causes within problem areas

An interesting problem area is the MRI department which, according to the hospital personnel, is an expensive resource, and which treated most patients only after two or more days after the application was filed, which indicates long waiting times. MRI treated 5457 patients in the period of 1-07-2010 to 30-06-2011. In figure 8 a throughput diagram is presented for all patients at MRI. How to read a throughput diagram is explained in chapter 4.1.3. Over time an increase in WIP is visible, the output stays stable but the input increases. The WIP shows a minimum of 153 applications waiting, the maximum is 410; the average is 282, and the median is 284 applications waiting.

Figure 9 shows that there is also an increase in throughput time. There is increasingly more time between the moment of application and the time of operation. The percentage of patients which are treated in 0 to 5 days stays the same over time but particularly the percentages of patients with very long throughput times increase strongly. The average throughput time is about 18 days. This can be explained by the following numbers: The minimum number of days between the application and operations is 0, the maximum 734, and the median 13. This means that the maximum of 734 days does not influence the average very much. In addition, I investigated the different types of operations at MRI. In all different types the increase in throughput time is visible. The number of each type of operation does not change remarkably. Furthermore, there is no considerable change in the number of operations performed for the different specialties.

(16)

Figure 8: Throughput diagram of applications and operations of the MRI department

A throughput diagram of all applications and operations at the MRI department in the period of 1-07-2010 to 30-06-2011. The data relate to the operations in the period of 2010 to April 2012.

Figure 9: Throughput time per month for MRI

(17)

4.2. Surgery patients at radiology

In the next subchapters the results of the diagnosis, using the framework of Soepenberg (2012), of the surgery patients of radiology are presented.

4.2.1. Step 1: Analyse the distribution of duration

In order to see the distribution of the duration for the surgery patients at the radiology department I made an overview of the time between the application and the operation of the activities for the surgical specialty. 1670 patients (22,62%) are treated the same day as the application, 615 patients (8.33%) are treated one day after the application. The remaining number of patients and the time between application and operation are shown in figure 10.

Also some peaks are visible around 90 days, 180 days and 365 days which are follow-up appointments. This is about the same as was found for the entire radiology department. This figure combines all kinds of surgery patients an all kinds of operations. However, it is possible that there are differences among subsets.

Figure 10: Time between application and operation – surgery patients

The number of days between the application of an operation and the operation itself for the surgery patients. The data relate to the number of patients in the period of 1-07-2010 to 30-06-2011.

4.2.2. Step 2: Analyse differences among subsets

(18)

4.2.2.1. Differences between radiology sub departments

In figure 11 the number of patients is shown for each radiology sub department. The Echo department treated the majority of the patients, Mammografie and CT also treated a lot of patients. OK treated only 1 patient. Differences in the number of patients are influenced by the different kind of treatments and technologies used in the different sub departments.

Figure 11: Number of surgery patients per sub department

The number of surgery patients per radiology sub department. The data relate to the number of patients in the period of 1-07-2010 to 30-06-2011. The sub departments have Dutch names: Angiografie for Angiography, CT for Computed Tomography, Dex for Dexa (bone density measurement), Doorlichting for Screening, Interventies for Interventions, Mammografie for Mammography, MRI for Magnetic Resonance Imaging, Nucl. Geneesk. for Nuclear Medicine, and OK for Operating Room.

(19)

Figure 12: Throughput time per sub department – surgery patients

Throughput time per radiology sub department, displayed in categories of the number of days (see the legend). The data relate to the number of patients in the period of 1-07-2010 to 30-06-2011. The sub departments have Dutch names: Angiografie for Angiography, CT for Computed Tomography, Dex for Dexa (bone density measurement), Doorlichting for Screening, Interventies for Interventions, Mammografie for Mammography, MRI for Magnetic Resonance Imaging, Nucl. Geneesk. for Nuclear Medicine, and OK for Operating Room.

4.2.2.2. Differences in urgent and less urgent patient groups

In order to investigate whether there are differences between urgent en less urgent patient groups I categorized the patients using primary diagnosis codes. Surgery patients in the categories Rectum, Colon, Longen and Mamma (Oncologie) should be treated within two weeks after they are placed on a waiting list. This means that when they need to be treated at the radiology department this should be possible within two weeks. For surgery patients in the category Traumatologie, urgency varies from a few days, within a week, and two weeks after placement on a waiting list. For surgery patients in the category Vaten, urgency varies from within one week, two weeks or elective.

(20)

When I compare these results to the planning matters described in the previous paragraph, at first sight there are no alarming numbers. However, there is a lot of variety in each category and I do not know the number of patients belonging to this variety. Therefore, it is hard to draw firm conclusions. For example, in the category Mamma, about 20 percent of the patients is not treated within two weeks but I do not know if this is bad or not bad.

Figure 13: Number of patients per primary diagnosis – surgery patients

The number of surgery patients per primary diagnosis. The data relate to the number of patients in the period of 1-07-2010 to 30-06-2011. The categories of primary diagnosis have Dutch names: Colon for Colon, Longen for Lung, Mamma for Breast, Rectum/Maag for Rectum/Stomach, Traumatologie for Traumatology, and Vaten for Vessels.

Figure 14: Throughput time per primary diagnosis – surgery patients

(21)

4.2.3. Step 3: Analyse differences over time

The next step in this diagnosis is analyzing time dependency. In figure 15 the throughput diagram is shown which contains cumulatively all operations with the corresponding applications. How to read a throughput diagram is explained in chapter 4.1.3. The curves run almost parallel upward even when I zoom in per month. The vertical distance between the input curve and the output curve depicts the work-in-progress (WIP). The WIP varies over time but not in remarkable quantities (10 percent or less from average). However, in mid-February the WIP starts dropping. This is partly due to a lack of data because the processed data continue until April 2012. This means that applications (and their corresponding operations) filed 9 months or more before April 2012 are not included in this throughput diagram. In figure 10 is shown that a significant part of the applications is filed 9 months or more before the operation takes place. Therefore, missing data may explain the decrease in WIP. In order to decrease these missing data, I tried to remove the follow-up appointments, but using the available data, it is not possible to do that. For radiology, I simply do not know which operations are follow-up appointments.

Figure 15: Throughput diagram of applications and operations at Radiology – surgery patients

A throughput diagram of all applications and operations of surgery patients at the department of radiology in the period of 1-07-2010 to 30-06-2011. The data relate to the operations in the period of 2010 to April 2012.

4.2.4. Step 4: Analyse the patient journey process

(22)

However, I do have some information available about the patient-journey-time-promising process based on the conversations I had at the hospital. During the conversations with the hospital management I learned that the waiting times communicated with the patients are based on estimates and may vary from week to week. During this estimation the hospital takes into account its capacity and patient preferences. For most treatments, there is no guidance with regard to the treatment times. So, the patient-journey-time-promising process is an unstandardized way of working. Further research needs to be done on this topic.

4.2.5. Part 2: Determining PPC causes within problem areas

An interesting problem area is the MRI. According to the hospital personnel, MRI is an expensive resource. In chapter 4.2.2.1 I showed that MRI treated most surgical patients only after two or more days after the application was filed. This is also the case for the arbitrary patients of the radiology department. Furthermore, in chapter 4.1.5 a trend is presented of increasingly more days between the application and operation for the entire radiology department. So, I investigated if this is also the case for surgery patients which are treated at MRI.

(23)

Figure 16: Throughput diagram of applications and operations of the MRI department – surgery patients A throughput diagram of all applications and operations of surgery patients at the MRI department in the period of 1-07-2010 to 30-06-2011. The data relate to the operations in the period of 2010 to April 2012.

Figure 17: Throughput time per month for MRI – surgery patients

(24)

5. Discussion

In this chapter the results of chapter 4 are discussed. From the perspective of the entire radiology department and from the surgery patients I showed that most patients are treated very fast. The application time is regularly 0 days, since patients can go from the consultation hours to radiology the same day, for almost all kinds of diagnoses. So, the general distribution of duration is perceived as good.

When investigating the different sub departments of radiology I found differences between the departments, but these differences can be explained by the nature of the treatments done there. However, the MRI department is characterized by a long duration between the application and operation. This is an expensive resource which treats a relatively large number of patients. Therefore, I investigated the MRI department more in-depth. From the perspective of the entire radiology department, the results show an increasing work-in-progress (WIP) over time concomitant with an increasing throughput time. This holds  but less clearly  for the surgery patients, which were investigated separately. One of the probable cause found for this phenomenon at MRI is the capacity which cannot handle the increasing input. Also, the conceptual model, presented in chapter 2.2, suggests that the available capacity can be one of the causes. When the input changes over time and the output stays the same because of lack of extra capacity, the throughput time can increase. Further in-depth research is needed in order to determine the actual production planning and control (PPC) causes.

Furthermore, the different specialties within the hospital were compared. The radiology department gets the most applications from the specialty surgery. The throughput time of surgery appears to be normal. The throughput time of the other specialties shows no noteworthy cases. Also, for the surgery patients urgent and less urgent patient groups were investigated. When comparing the results to the urgency matters of the different groups I did not find any alarming numbers.

In the next step time dependency was investigated. The input, WIP and output show no remarkable changes over time for the entire radiology department and neither did it for the surgery patients.

(25)

6. Conclusion

In this Master thesis a diagnosis was made of delaying factors in the care path of surgery patients, specifically surgery patients who are treated at the radiology department at the hospital. The objective of this research is to help the hospital management with improving the processes in the surgical department. The following research question was investigated: “Which factors are the main factors of throughput time delays in the care path of patients in the surgical department of the hospital?”. To answer this question a literature review was done which revealed that the framework to diagnose delivery reliability in make-to-order (MTO) companies of Soepenberg et al. (2012) is the only usable model for doing the diagnosis. However, the model had to be adapted to hospital environments. This is possible and was elaborated in chapter 3. A conceptual model was made using literature and information learned during conversations and interviews at the hospital. The conceptual model shows that the available capacity in the hospital affects the processes which finally affect the throughput time. The methods used for this research are presented in chapter 3. Chapter 4 presents the results of diagnosis; these results are discussed in chapter 5.

In answer to the research question, the conclusion can be made that there is only one factor which causes throughput time delays in the care path of patients: the MRI department. The available capacity appears to be the one of the most probable cause for the increase of throughput times when the input increases.

Research implications and recommendations

Because of time limitations, most analyses in this research are done on a high level of aggregation. If some analyses had been continued on a lower aggregation level they could have revealed more results.

(26)

7. References

Cassidy-Smith, T.N., Baumann, B.M., and Boudreaux, E.D. (2007). The disconfirmation paradigm: Throughput times and emergency department patient satisfaction. Journal of Emergency Medicine, 32, 1, 7-13.

Ekelund, U., Kurland, L., Eklund, F., Torkki, P., Letterstal, A., Lindmarker, P., and Castrén, M. ( 2011). Patient throughput times and inflow patterns in Swedish emergency departments. A basis for answer, A National Swedish Emergency Registry. Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine, DOI: 10.1186/1757-7241-19-37.

Ling, L., and Benton, W.C. (2003). Hospital capacity management decisions: Emphasis on cost control and quality enhancement. European Journal of Operations Research, 146, 596-614.

Ringard, A., and Hagen, T.P. (2011). Are waiting times for hospital admissions affected by patients’ choices an mobility?. BMC Health Services Research, DOI: 10.1186/1472-6963-11-170.

Siciliani, L., and Hurst, J. (2004). Explaining waiting-time variations for elective surgery across OECD countries. OECD Economic Studies, 1, 95-123.

Soepenberg, G.D., Land, M.J., and Gaalman, G. (2008). The order progress diagram: A supportive tool for diagnosing delivery reliability performance in make-to-order companies. International Journal of Production Economics, 112, 1, 495-503.

Soepenberg, G.D., Land, M.J., and Gaalman, G.J.C. (2012). A framework for diagnosing the delivery reliability of make-to-order companies. International Journal of Production Research, DOI: 10.1080/00207543.2011643251.