Reduce waiting for a diagnostic service:

(1)

Reduce waiting for a diagnostic service:

Exploring demand and supply characteristics in a nuclear

therapeutic and diagnostic clinic in the Netherlands

(2)

Emma Zijlstra | MBA Operations & Supply Chains

2 Reduce waiting for a diagnostic service:

Exploring demand and supply characteristics in a nuclear

therapeutic and diagnostic clinic in the Netherlands

-August 26, 2011 -

Emma Zijlstra

Studentnumber: 1884697

Star Numanstraat 103B

9714 JN Groningen

tel.: +31 (0)638223588

e-mail: emma_zijlstra@hotmail.com

University of Groningen

Faculty of Economics and Business

MSc Business Administration

Specialization Operations & Supply Chains

Supervisor/ University of Groningen

Dr. M.P. Mobach

(3)

3 A

BSTRACT

Purpose – The purpose of this study is to explore the capacity planning and control of a diagnostic clinic in order to increase speed and the level of utilization with respect to the current capacity level. Specifically the influence of patient categorization, demand variability, work design and supply variability on capacity planning and control is investigated.

Design/ methodology/ approach – The current study was made in the context of discovery and exemplification. With time study preparation times, set-up times, cool down times, scan times, and scan room times of 75 patients were analyzed. The care related activities of nuclear workers are analyzed by applying the work sampling method which included 977 frequencies for analysis.

Findings – FDG wholebody examinations take less time than is actually planned and the weight of patients significantly influence the examination times. Heavy weighted patients costs more camera time than low weighted patients. Nuclear workers spend on average 34% on care related activities like preparing a patient, scanning a patient, picking-up patient, data processing and reading the patients record. Surprisingly, the majority of time is spend on other activities, frequently on administration, discussions, and own time.

Practical implications – Managers of diagnostic imaging centers can actively use capacity planning and control as a tool to reduce the waiting times and increase the utilization levels with respect to the current capacity level. By specializing tasks of nuclear worker, scheduling breaks and preparation times in advance and apply realistic examination times for scheduling based on weight groups the operational performance can be significantly improved.

Originality/ value – The results of this paper contribute to an efficient use of the current capacity and to more promptly diagnostic imaging which is essential for cancer patients for optimal treatment.

(4)

4 P

RE

-

FACE

This master thesis is a result of the study during my graduation for the Master Operations & Supply Chains at the University of Groningen, conducted at the University Medical Center Groningen (UMCG). Using this opportunity, I would like to thank all the people who supported me inside and outside the hospital that made this research successful.

First of all I would like to thank my internal supervisor, Ir. Peer Goudswaard, for all the time and effort he has devoted to me. He has given me the opportunity to do research within the UMCG and supported me during this research when needed. This wonderful and rewarding time would not have been possible without him. Next to Peer, I would like to thank the operations manager of the NMMI clinic, Annie van Zanten, for making this project possible at the NMMI clinic and for the support during this study.

Besides the internal support I would like to thank my supervisor from the University of Groningen, Dr. Mark Mobach. It was the second time we cooperated during a research project and I have perceived it just as pleasant and successful as the first time. I want to thank him for the time and effort he has devoted to me and for providing useful feedback. It was really nice and motivating to experience so much confidence and interest in my work. Without your help I would not have been able to present this final result. I would also like to thank the second assessor from the University of Groningen, Prof.dr.ir. Kees Ahaus, for evaluating this thesis.

Furthermore, I would like to thank the two staff employees of Healthcare Operations Management, Ir. Tjibbe Hoogstins and Ir. Igor van der Weijde, for the support during this study and for being my companions at the office. Unforgettable was the invitation for the UMCG volleyball tournament which has been a unique experience.

And last but not least a lot of people inside the hospital helped me with understanding the complex processes of the NMMI clinic. I really would thank them all for the time they invested in me and showing me around. Without exception they were always ready to help me. In particular I would like to thank Paul Snick who familiarized me with the PET scans and the issues surrounding the PET scans. Next to him I would like to thank the manager administration, Peggy Radjiman, and the manager nuclear workers, Hans ter Veen, for all the nice conversations and for the time to answer all my questions.

Thanks all!

Emma Zijlstra

(5)

5 T

ABLE OF CONTENTS

Chapter 1: Introduction ... 8

1.1 The UMCG ... 8

1.2 The Nuclear Medicine & Molecular Imaging outpatient clinic ... 9

1.3 Content of this research ... 10

Chapter 2: Research Design ... 11

2.1 Introduction ... 11

2.2 Background and motivation of this research ... 11

2.3 Problem statement ... 12

2.4 Conceptual model and sub-questions ... 12

2.5 Operationalization ... 15

Chapter 3: Theoretical framework ... 16

3.1 Introduction ... 16

3.2 Independent variables ... 17

3.2.1 Demand ... 17

3.2.2 Supply ... 19

3.3 Dependent variables ... 23

3.3.1 Capacity planning and control ... 23

3.3.2 Operational performance ... 28

3.4 Conclusion theoretical framework ... 32

Chapter 4: Methodology ... 34 4.1 Research strategy ... 34 4.2 Data analysis ... 35 4.2.1 Archiving data ... 35 4.2.2 Observations ... 36 4.2.3 Interviews... 38 Chapter 5: Results ... 39 5.1 Results demand ... 39 5.1.1 Patient categorization ... 39 5.1.2 Demand variability ... 41 5.2 Results supply ... 43 5.2.1 Work design ... 43 5.2.2 Supply variability ... 44

5.3 Results capacity planning and control ... 46

5.3.1 Resource allocation ... 46

5.3.2 Planning and control activities ... 47

5.4 Results operational performance ... 51

5.4.1 Speed ... 51

5.4.2 Utilization ... 52

(6)

6

Chapter 6: Discussion ... 57 6.1 Demand ... 57 6.1.1 Patient categorization ... 57 6.1.2 Demand variability ... 57 6.2 Supply ... 58 6.2.1 Work design ... 58 6.2.2 Supply variability ... 58

6.3 Capacity planning and control ... 59

6.3.1 Resource allocation ... 59

6.3.2 Planning and control activities ... 59

6.4 Operational performance ... 60 6.4.1 Speed ... 60 6.4.2 Utilization ... 60 Chapter 7: Conclusion ... 62 7.1 Conclusions ... 62 7.2 Recommendations ... 62 7.2.1 Demand ... 62 7.2.2 Supply ... 63

7.2.3 Capacity planning and control ... 64

7.2.4 Operational performance ... 64

7.3 Further research ... 65

References ... 66

Appendix I Search strategy... 70

Appendix II Work sampling ... 71

(1) Calculation amount of observations ... 71

(2) Classification activities ... 72

Appendix III Interviews... 74

(1) Manager medical records administration... 74

(2) Manager nuclear workers ... 74

(3) Nuclear workers ... 75

Appendix IV Results Demand ... 77

(1) Histogram of weight ... 77

(2) Distribution planned examination (weekly based) ... 77

(3) Histogram of demand ... 78

(4) Probabilistic estimation demand ... 79

Appendix V Results Supply ... 80

(1) Preparation time ... 80

(2) Set-up time ... 80

(3) Cool down time ... 81

(4) Scan time ... 82

(5) Scan room time ... 85

(7)

7

Appendix VI Results capacity planning & control ... 89

(1) Resource allocation... 89

(2) Loading... 90

(3) Scheduling ... 91

Appendix VII Results Operational Performance ... 93

(1) Reducing days ... 93 (2) Waiting times ... 94 (3) Opening hours ... 95 (4) Machine utilization HR+ ... 96 (5) Machine utilization mCT ... 97 (6) Estimated utilization ... 97

(8)

8 C

HAPTER

1:

I

NTRODUCTION

This research was initiated by the management of the ‘healthcare operations management’ department and the management of ‘Nuclear Medicine and Molecular Imaging’ (NMMI) of the University Medical Center Groningen (UMCG). The healthcare operations management department aims to improve the logistical processes of healthcare in the UMCG.

The management of the NMMI clinic mentioned a capacity planning and control problem because resources are not optimal used. The NMMI clinic deposits this problem at the staff department ‘healthcare operations management’. Nowadays, costs have to decrease in the hospital because of healthcare cuts while the quality has to increase. This is a challenging task for hospitals and for this reason research to efficiency is necessary.

This chapter includes first the background of the organization followed by a description of the particular diagnostic clinic. Finally, the content of this study will be elaborated.

1.1 T

HE

UMCG

The UMCG is one of the biggest hospitals of the Netherlands and the largest employer of Northern Netherlands. The UCMG is the result of a merger between the Academic Hospital Groningen (AZG) and the Faculty of Medical Sciences. Patients attend the UMCG for basic care but also for specialist diagnoses, examinations and treatments. Every day around 18.000 people enter the UMCG consisting staff, patients, visitors and students.

The UMCG has more than 10.000 employees, more than 1.300 beds, more than 35.000 patient admissions a year, almost 500.000 outpatient visits and more than 300.000 nursing days.

The organization structure is decentralized (figure 1.1). Every sector contains several departments, care facilities and a business office. This study was initiated under supervision of the staff department ‘healthcare operations management’ and is part of the business office of sector E and has a support staff function.

(9)

9 1.2 T

HE

N

UCLEAR

M

EDICINE

&

M

OLECULAR

I

MAGING OUTPATIENT CLINIC

The NMMI diagnostic clinic aspires to perform patient care at the highest possible value (Mission, 2009). The NMMI clinic is part of sector E (figure 1.1). The NMMI clinic concerns the development, improvement, and application of radiopharmaceuticals for functional imaging. The aim of the clinic is an optimal therapy with open sources of radioactivity in combination with quantitative diagnostic techniques. The clinic also acquires scientific information on body function in experimental animals and healthy volunteers. The study is focused on patient care. The service the NMMI clinic delivers to patients is: ‘make various disorders and metabolic processes visible by injecting radiopharmaceuticals in the bloodstream of a patient’. The NMMI clinic makes this visible by using different types of cameras. Patients are referred to the NMMI clinic by a referring specialist. The results of the research on the NMMI clinic support the specialist to diagnose and treat the patient.

This clinic contains five types of cameras and in total contains seven examination rooms. Yearly based there are 14.500 treatments and around 180 therapies. The total work size of the NMMI clinic is around 55 employees (50 fte) and around 20 PhD students. Staff exists out of nuclear doctors, nuclear workers (laboratory technicians), medical physicists, radio chemists, biologists and pharmacists (figure 1.2). Daily, around 40 employees are present at the clinic.

The NMMI clinic can be distinguished in the PET side and the SPECT side. PET stands for ‘Positron Emission Tomography’. This is an imaging technique which makes use of three dimensional images of functional processes in the body. SPECT stands for ‘Single photon Emission Computed Tomography’ and makes use of the ‘gamma ray’ technique which makes two dimensional images from multiple angles. The PET cameras provide higher resolution images and make use of short-living radiopharmaceuticals in comparison with the SPECT cameras. The PET images are popular for the referring specialists due to higher resolution images the specialists are able to diagnose and treat a patient better. The NMMI clinic contains two PET cameras. One camera can only provide PET scans and the second camera can provide PET/CT scans. A CT scan stands for ‘computed tomography’ and provide images of the cross section of the body (Van der Velde et al., 2011). According to Van der Velde et al. (2005) a PET scan can visualize the tumor metabolism. The combination of those two techniques is called a PET/CT scan and provides accurately registered anatomical localization of structures seen in the PET scan (Kinahan et al., 1998). An important benefit from the patients’ points of view is that only one visit is necessary instead of two separate visits for a PET and CT scan. The differences between a CT scan, a PET scan and a PET/CT scan are shown in figure 1.3 on the next page.

(10)

10

FIGURE 1.3: DIFFERENCES IMAGING CT SCAN, PET SCAN, AND PET/CT SCAN

The NMMI clinic contains two PET cameras and four SPECT cameras. Because of the limited number of PET cameras and the expected growing demand of PET images due to an aging population this study will focus on the PET side.

1.3 C

ONTENT OF THIS RESEARCH

This thesis is organized as follows. This paper starts with the design of this research in chapter 2. In chapter 3 the theoretical framework describes the related literature which spans demand, supply, capacity planning and control, and operational performance. In chapter 4 the research strategy and data analysis is described. Chapter 5 will describe the ‘results’ of this research and chapter 6 the ‘discussion’ of the results related to the literature framework. Finally, in chapter 7 ‘Conclusion’, the most important conclusions, the recommendations and present avenues for future research are presented.

Brain

(11)

11 C

HAPTER

2:

R

ESEARCH

D

ESIGN

2.1 I

NTRODUCTION

In order to write a scientific report this research design is conducted based on the research process of Welker and Broekhuis (2010). Welker and Broekhuis (2010) developed a research process. The research design is part of it and elaborates the process from problem statement to research design. This chapter will first elaborate the background of the problem and motivation of the research (2.2) followed by the problem statement (2.3). Chapter 2.4 will discuss the conceptual model and the sub-questions of this research. Finally, the conceptual model will be operationalized in chapter 2.5.

2.2 B

ACKGROUND AND MOTIVATION OF THIS RESEARCH

A graduation has to be completed in order to finish the Master Business Administration in Operations & Supply Chains. As mentioned in the first chapter this research is conducted for the University Medical Center Groningen (UMCG). Hospitals have the challenge to decrease costs while improving quality. The manager of the NMMI clinic wants to decrease costs and improve quality more specifically for the PET scans.

The NMMI diagnostic clinic on the PET side consists of one waiting room, two examination rooms, two preparation rooms, one relaxation room and five nuclear workers a day. The route that patients take is shown in figure 2.1.

After arriving, the patient has to register at the NMMI desk first. After registering the patient has to wait in the waiting room. The patient is picked up by the nuclear worker and gets prepared in the preparation room. After being prepared the patient goes to the relaxation room. When this time is elapsed in the relaxation room images can be constructed from the body in the examination room. After imaging the body the patient can leave the clinic.

The PET-scans can be performed on two cameras manned by several nuclear workers (laboratory technicians). The two cameras are called the mCT and the HR+. The main difference between those cameras are the technological capabilities; the HR+ camera can make a PET scan, however these scans are inferior to the mCT camera because the mCT camera can provide better images (higher resolution) and an extra CT scan. Therefore demand increases for the mCT camera and “PET/CT gains more and more importance in oncology imaging” (Boellaard et al., 2010, pp. 183) which lead to longer waiting times. According to Spiro et al. (2008) early diagnostics such as PET are essential for cancer patients for optimal treatment.

The scarce resource problem for PET scans results in longer waiting times for patients which results in high labor costs because the NMMI clinic expanded the opening hours. The NMMI clinic defined different performance indicators to improve quality. The current performance indicators and objectives which are relevant for this study are:

- Throughput time between the examination request and patient contact (examination) – 80% within 5 working days

- Utilization level cameras – more than 70%

These performance indicators are part of the EFQM model and are already defined during the year 2010. This study will be focused to analyze the current performance of these indicators and improve the throughput time and the utilization level. To analyze the scarce resource problem from the external perspective the

(12)

12

operations manager of the NMMI clinic asked the supported service ‘Healthcare Operations Management’ of UMCG for advice. Under supervision by the manager ‘Healthcare Operations Management’ this research is conducted. Ultimately, this thesis should provide the NMMI clinic a better understanding about capacity planning of the resources and the consequences for performance of the NMMI clinic. This study should provide faster and therefore better care to the patients of the NMMI clinic.

2.3 P

ROBLEM STATEMENT

As mentioned in the previous paragraph, the management noticed a problem in capacity planning and control. Therefore the following research objective is formulated.

To reach this goal the following research question has to be answered:

To optimize the planning the first step to do is finding out which factors are influencing the planning. After analyzing the factors which are currently influencing the capacity planning of the department an improvement step can be made by analyzing how this actual situation can be optimized.

2.4 C

ONCEPTUAL MODEL AND SUB

-

QUESTIONS

The central function of a hospital is to provide patient care (Vissers and Beech, 2005). Nowadays, hospitals have the challenge to decrease costs because of the health care cuts while the quality of patient care has to increase. According to McLaughlin and Hays (2008) many answers related to this lies within organizational operations. Effectiveness and efficiency can be increased by operational improvements; how to deliver high-quality care in a consistent, efficient manner (McLaughlin and Hays, 2008). To be able to improve processes it is important to firstly know the process and its desired inputs and outputs (McLaughlin and Hays, 2008). According to Vissers and Beech (2005) operations management refers to the planning and control of the processes that transform inputs into outputs.

The inputs in the healthcare process of the NMMI clinic are the patients, the referring specialists, the radiopharmaceuticals, the cameras and the nuclear workers (laboratory technicians). The output is that patients are examined. This transformation process is shown in figure 2.2 (Krajewski and Ritzman, 2005).

FIGURE 2.2: TRANSFORMATION PROCESS

Analyze and optimize the capacity planning of the NMMI clinic to be able to decrease waiting times and optimize the utilization level with respect to the current capacity.

(13)

13

Because hospitals are not dealing with products but with services it is not possible to hold inventory. Inventory in a service is a waiting customer (Fitzsimmons and Fitzsimmons, 2006) therefore it is vital to make good use of fixed assets (Nordgren, 2009). Planning and control brings two entities together, namely supply and demand. Allocation of resources such as cameras, preparation facilities and nuclear workers is an important issue for effective and efficient operations management (Vissers and Beech, 2005). When an examination is delayed and patients have to wait patients can face substantial risks of complication or even death (Wang, 2004). This particularly applies to patients for PET scans. According to Boellaard et al. (2010) the most common PET examinations are for detecting, staging, re-staging as well as for assessment of therapy response of oncology patients. This can be explained by the fact that in the Netherlands the cause of death by cancer is relatively still growing (Gezondheid en Welzijn, CBS).

The objective of this thesis refers to the influence of capacity planning and control on the operational performance (waiting time and utilization level). According to Slack et al. (2004) the definition of capacity planning and control is:

According to Vissers and Beech (2005) a capacity planning problem can be caused by three things, namely (1) resource availability not in balance with the demand, (2) timing of allocations in terms of periods that a resource is available may lead to peaks and troughs in the workload, and (3) the capacities of different resources that are required simultaneously are not balanced, resulting in bottlenecks or under-utilization. According to Vissers and Beech (2005) the underlying reason for capacity loss is that allocation of resources often tends to be based on historical rights instead of on the requirements for resources resulting from the flow of patients. This can be caused by lack of availability of procedures and methods that can be used to regularly update resource allocations (Vissers and Beech, 2005). When allocations are not based on the flow of patients it can easily result in less optimal use of resources which results in fewer patients examined.

Different capacity strategies are developed to set the effective capacity of the operation so that it can respond on the demand: Long-, medium-, and short-term capacity planning and control (Slack et al., 2004). The long-term planning and control (months/years) contains plans about what they intend to do, what resources are needed and what objectives the organization hopes to achieve. This long-term planning will make use of forecasts of demand in aggregated terms. According to Fitzsimmons and Fitzsimmons (2006) the aim of strategic capacity planning (long-term) is to determine the suitable level of service capacity by specifying the proper mix of facilities, equipment, and labor that is required to meet estimated demand.

The medium-term planning and control (days/weeks/months) contains more detailed information. A characteristic of medium-term planning and control is that partially disaggregated demand forecasts can be used (Slack et al., 2004). This can be done by distinguishing different types of demand. In a medium-term planning resources will also be set at a disaggregated level, like different categories of staff. In short-term planning and control (hours/days) many resources are not changeable anymore at large-scale. Interventions and changes are on an ad hoc basis.

This research will be focused on the medium-term planning and control. Resources and objectives are already available at the NMMI clinic (long-term planning) and now it is necessary to analyze the medium-term planning to look at the planning in a partially disaggregated manner. More specifically on the capacity planning and control of the PET scans to be able to reduce costs and improve quality. According to Vissers and Beech (2005) this is the tactical level (resources planning and control).

(14)

14

Based on the literature mentioned above the conceptual model is developed and is shown in figure 2.3 on the next page. This conceptual model forms a framework for this research to be able to answer the main question:

The dependent variable is ‘operational performance’ and means how to deliver high-quality care in a consistent, efficient manner (McLaughlin and Hays, 2008). The mediating variable ‘capacity planning and control’ is determined by the reconciliation between ‘supply and demand’ and positively influence operational performance. Supply and demand are the two independent variables. Smoothing demand and supply can positively influence the mediator ‘capacity planning and control’ and therefore can positively influence the operational performance.

FIGURE 2.3: CONCEPTUAL MODEL

In order to answer the main research question several sub-questions are formulated. These sub-questions are derived from the several concepts of the conceptual model shown in figure 2.3:

The next sub-questions are related to the demand concept.

1) Which factors of demand influence the capacity planning and control?

2) What is the actual situation of demand characteristics which influence capacity planning and control?

The following sub-questions are related to the supply concept.

3) Which factors of supply influence the capacity planning and control?

4) What is the actual situation of supply characteristics which influence capacity planning and control?

The way in which capacity planning and control is organized is determined by the reconciliation of demand and supply. The following sub-questions are related to the ‘capacity planning and control’ concept.

5) How are resources allocated?

6) What is the actual situation of the capacity planning and control?

Finally, the following sub-questions are related to the concept ‘operational performance.’ 7) What is the current level of performance in terms of speed?

8) What is the current level of performance in terms of utilization?

(15)

15 2.5 O

PERATIONALIZATION

Concept Operationalization

Demand Patient categorization

Demand variability

Supply Work design

Supply variability Capacity planning and control Resource allocation

Planning and control activities - Loading

- Sequencing - Scheduling

- Monitoring and control Operational performance Speed

- Waiting times Utilization

- Machine utilization - Labor utilization

TABLE 2.1: OPERATIONALIZATION

(16)

16 C

HAPTER

3:

T

HEORETICAL FRAMEWORK

The result of chapter 2 was an operationalization of the conceptual model. This theoretical framework will further elaborate these operational concepts. Chapter 3.1 will first discuss a short introduction. Followed by chapter 3.2 that elaborately describes the independent variables ‘demand’ and ‘supply’ and how these variables can influence the dependent variables. Chapter 3.3 discusses the dependent variables ‘capacity planning and control’ and ‘operational performance’. Finally in chapter 3.4 the conclusion of the theoretical framework will be elaborated.

3.1 I

NTRODUCTION

According to Mango and Shapiro (2001) the flow of patients through a hospital is limited by the weakest link of the hospital. For example, if too few nuclear workers or too few cameras’ are available at the NMMI clinic, patients have to wait longer before they can be examined at the NMMI clinic. This problem will turn in delay for the next patients. Therefore developing an efficient process is important. An efficient process in the healthcare industry context means a process which reduces costs and improves quality (Cayirli and Veral, 2003). For that reason it is important to identify the potential bottlenecks and alleviate them by balancing the process, smoothing demand and control capacity (Mango and Shapiro, 2001). According to Mango and Shapiro (2001) when the problems becomes clear it is usually possible to eliminate these problems by applying a better planning. According to Slack et al. (2004, p. 325) planning can be defined as ‘a formalization of what is intended to happen at some time in the future’. According to Cayirli and Veral (2003) a well designed planning creates the ability to increase the utilization of expensive employees and equipment-based medical resources. To develop an efficient process on the PET side of the NMMI clinic it is essential to do this literature study. In the Netherlands the population aged over 65 is increasing. This growth will be at the maximum around the year 2040. During those 30 years the population of elderly will grow from 2.5 million to 4.6 million (Bevolkingsprognose 2010-2060). This demographic change will increase the overall demands of health care and it is likely that it will affect the ways in which health care is delivered. Because of the increasing demand services have to be modernized in such a way that they can make more intensive and efficient use of existing health care resources (Vissers and Beech, 2005). This literature study will first consider theories about the independent variables ‘demand’ and ‘supply’ followed by theories about the dependent variables ‘capacity planning and control’ and ‘operational performance’ (figure 3.1).

(17)

17

3.2 I

NDEPENDENT VARIABLES

3.2.1 D

EMAND

The first independent variable is demand. On the demand side an operation has customers (Slack et al., 2004). The customers of a diagnostic clinic are referring specialists and patients. According to Green et al. (2006) medical imaging facilities are assessed by a wide range of patients, both inside and outside of the hospital.

According to Fitzsimmons and Fitzsimmons (2006) smoothing demand can reduce the required capacity, can utilize capacity which result is improved productivity. In the first section of this subparagraph patient classification is discussed which can influence the allocation of resources. And the second section discusses the demand variability.

3.2.1.1 P

ATIENT CATEGORIZATION

In the healthcare industry various patient types exist. Several studies assessed how patient characteristics can influence the allocation of resources. However, the majority of studies assume patients are homogeneous and are scheduled on a first in first out (FIFO) method (Caryirli and Veral, 2003). Walter (1973) found that even simple grouping of patients already results in improvements in doctors’ idle time. According to Vissers and Beech (2005) to make a planning more manageable the variety of patients need to be categorized. The underlying assumption is that the patient population can be classified into groups based on examination time characteristics. In practice this implies the scheduler to allocate patients to appropriate slots reserved for each patient type (Cayirli et al., 2008).

According to Green et al. (2006) patients for medical diagnostic facilities can be grouped into three broad categories: (1) outpatients, (2) inpatients, and (3) emergency patients. Walter (1973) faced a substantial improvement in terms of doctors’ idle time by simply grouping ‘outpatients’ and ‘inpatients’. Other suggestions for groups are patients age, physical mobility, and type of service. He found that variables like age physical mobility and type of service influence examination times. On average, older patients with limited mobility (trolley or wheelchair) require more service time than the younger and walking patients. Therefore only a distinction between inpatients and outpatients is too limited. Walking inpatients require shorter service times than outpatients who need a trolley or wheelchair. According to Cayirli et al. (2008) the implementation of patient categorization results in improving patient flow times, queue lengths, and doctor utilization.

3.2.1.2 D

EMAND VARIABILITY

According to Vissers and Beech (2005) variability in demand can affect the capacity requirements. Variability in demand for service need not be accepted as inevitable. Demand variability can be distinguished as follows:

Variability in arrival times (Slack et al., 2004) Variability in types of examinations (Litvak, 2000)

Variability in arrival times

According to Slack et al. (2004) no variation in arrival times rarely occurs. Variability in demand for hospital services can have major effects on hospital costs (Baker et al., 2004). Service systems need to smooth their demand (Fitzsimmons and Fitzsimmons, 2006) otherwise high variability hospitals need more capacity to meet high demand peaks (Baker et al., 2004). Fluctuations in demand can occur seasonally, weekly, and daily. Seasonal fluctuations in demand are brought about by factors such as weather, holidays, and vacation periods (Schönsleben, 2007).

(18)

18

and appointment decisions have to be made at the same time. In case of this the mapped arrival process is the same as the actual arrival process. However, appointments for specialty care clinics are booked by the referring physician (or clinical assistant) at the end of each session. That means that appointment requests are accumulated which means that the mapped process has batch arrivals that occur at regular intervals (Gupta and Denton, 2008). For example, at the end of each day the referring physicians send the total of requests of the day to the PET administration.

Variability in types of examinations

Besides variability in arrival times, outpatient clinics have to deal with variability in types of demand, more specifically types of examinations. When patients were homogeneous, they all had the same disease, the same degree of illness, and the same response to therapy (Litvak, 2000). This is not the reality, all patients differ. Therefore the challenge for a healthcare system is to create of a naturally variable group of patients a homogenous group to achieve higher efficiency (Litvak, 2000). According to Vissers and Beech (2005) patients have to be grouped according to their utilization of resources. This means that patients in the same group have a similar length of stay and require on average the same amount of nursing and operating time.

Simulation is a used method for evaluating the influence of the use of patient categorization on variability of consultation times (Klassen and Rohleder, 1996). The study of Klassen and Rohleder (1996) was conducted for family practice outpatient clinics and the result of the simulation study was that scheduling low variance patients in the beginning of the session resulted in the best performance (waiting time and idle time).

3.2.1.3 S

UB CONCLUSION

The first sub-question of this study related to the concept demand was: Which factors of demand influence the capacity planning and control? The answer on this question is demand categorization and demand variability.

According to Walter (1973) even simply categorization of patients results in decreased doctors’ idle time. By categorizing patients based on examination times the scheduler is able to schedule patients to appropriate slots (Cayirli et al., 2008). When appropriate slots are defined for the different patient groups the next step is to schedule the low variance patients group in the beginning of the session. This is the group where the standard deviation of the examinations times is the lowest.

Of course this is easier to say than to apply, because not always all groups of patients are present to schedule. According to Vissers and Beech (2005) variability in demand can affect the capacity requirements. A distinction is made between variability in arrival times and variability in types of examinations.

(19)

19 3.2.2 S

UPPLY

“Due to the fact health care services cannot be stored, it is vital to make good use of fixed assets” (Nordgren, 2009, pp. 123). Financial pressures on hospitals increase nowadays which make capacity management difficult (Jack and Powers, 2009). First, work design will be discussed. Secondly, theory about supply variability will be explored.

3.2.2.1 W

ORK

D

ESIGN

When humans participate in the tasks that are performed there are two basic categories of work systems: (1) manual work system, (2) worker-machine system (Groover, 2007). A third work system is the automated work system. Automation is the technology without human assistance for a given process. The distinction between worker-machine and automation is not always directly clear because many worker-machine systems operate with some degree of automation. Therefore automated work systems fall within the worker-machine system. The manual work system and worker-machine system will be further elaborated. This section will be followed by literature about redesigning work systems.

Manual work system

Engaging the human body to accomplish a task without using an external source of power is called manual work. Manual work can be subdivided in: (1) pure manual work, and (2) manual work using hand tools (Groover, 2007). Examples of pure manual work are an office worker filing documents in a file cabinet, or a dealer at a casino table dealing cards or a nurse moving a patient from one room to another. The common characteristic of pure manual work is that they consist of moving things. However manual tasks are commonly facilitated by the use of hand tools. Hand tools are small tools that are operated by the strength and the skills of the human. For example a general practitioner who is using a stethoscope to listen to the heart sounds of a patient.

Gilbreth and Gilbreth (1920) already studied the motion of work in the early 20th century. This study was focused on placing handicapped people after the First World War. The goal was to train handicapped with least waste and an increased production by standardizing motions. They state: “achieve the ‘one best way’ with the least unnecessary fatigue” which refers to standardizing processes to be able to increase efficiency.

Worker-machine system

The most widely used work system is a worker-machine system (Groover, 2007). This is when a worker operates powered equipment. The strengths and capabilities of a worker and a machine are combined and the result is synergistic. Figure 3.2 shows the interactions in a worker-machine system. The human controls the machine operation, and the machine displays the actions.

FIGURE 3.2: WORKER-MACHINE SYSTEM (GROOVER, 2007)

(20)

20

For the worker it starts with sensing. The five basic human senses are vision, hearing, touch, taste, and smell. These senses make a worker aware of where they are and what is happening. This is followed by information processing which can be considered as the workers brain. From the human senses it stores data in memory. The data will be operated by thinking, planning, calculating, solving problems, and making decisions about what actions to take. These actions will be taken with the human effectors which are fingers, hands, feet, and voice. These three human components will interact with the machine components in the worker-machine system. The worker will interact with the worker-machine controls that can be operated by the human effectors. Than the machine will process and this is followed by displaying the performance of the process. For example at the NMMI clinic the camera makes images (process) and the results of these images are shown on the computer (displays). Than the nuclear workers senses that the scan is finished and decides (information processing) to reposition the patient (action).

Work design

According to Campion et al. (2005) work design has a practical significance to organizational efficiency and satisfaction. For redesigning work systems four approaches have to be considered, namely mechanistic, motivational, perceptual, and biological. These models are linked with discipline base, recommendations, typical benefits and typical costs. The first approach is the mechanistic model and is drawn from the industrial engineering discipline. This model recommends specialization, simplification, and repetition. The benefits of this model are efficiency, easier staffing, and reduced training. The opposite is that satisfaction and motivation decrease. According to Groover (2007) this approach is often viewed negatively because jobs tend to be routine, boring, unchallenging, and unrewarding. Compensating these negative views can be done by three methods (Groover, 2007): (1) job enlargement, (2) job enrichment, and (3) job rotation.

Job enlargement refers to an increase in the number of activities included in the work. For example if a worker first fulfills only two steps of the process of patient and the job scope was expanded to completing the whole process for one patient. After job enlargement, the worker will experience the job as more interesting because of the variety of tasks.

Job enrichment refers to a greater responsibility of a worker. The work assumes greater responsibility for planning the work content and for inspecting the resulting work unit. Suppose a worker’s current job is simply positioning the patient, scan the patient, and reposition the patient. Someone else sets the machine up for the next patient, and someone else inspects the finished images. An example of job enrichment is to increase the scope of the job so that the worker is responsible for all different tasks.

The third method to eliminate the negative view of specialization is job rotation. A worker is trained to perform several different tasks. The term cross training is also used; a worker is trained to perform multiple tasks. The job becomes more interesting when a worker has the opportunity to perform a variety of tasks.

These three methods can prevent negative views of specializing tasks for the NMMI clinic to be able to work highly efficient and productive.

The second approach of Campion et al. (2005) is the motivational model. This model is drawn from the organizational physiology discipline. Variety, autonomy, and participation are recommendations from this model. These recommendations lead to satisfaction, intrinsic motivation, retention, and customer service. However the costs of this model are training, errors, and stress. This model usually provides ‘job enrichment’ as mentioned above (Campion et al., 2005).

The perceptual model is the third approach mentioned by Campion et al. (2005) and is drawn from human factors and the experimental psychology approach. This model is concerned with reducing the information-processing requirements. The benefits of this model are reduced errors, fewer accidents, and less mental overload. The disadvantages of this model are that they can lead to boredom and monotony.

(21)

21

requirements and reducing environmental stressors in order to increase physical comfort and reduce physical stress and fatigue. The opposite of the benefits are the financial costs and inactivity.

As mentioned above the different approaches have advantages and disadvantages and therefore a trade-off has to be made. For this study the mechanical model will fit with the objective of this study to provide a high efficiency.

3.2.2.2 S

UPPLY VARIABILITY

Variability also occurs in work systems. According to Hopp and Spearman (2001) variability can be defined as follow:

This definition means for example a group of individuals who all have the same weight have no variability in weight. In manufacturing there are many attributes in which variability is of interest: physical dimensions, process times, machine failure/repair times, quality measures, temperatures, material hardness, setup times, and so on. According to Mango and Shapiro (2001) variability on the supply side in hospitals often exist because of poor deployment of the resources. For example inconsistencies in times that doctors need or poor estimates how long examinations take. According to Cayirli and Veral (2003) most studies assume that patient are homogenous for scheduling purposes and use identically service times for all patients. When a high variability of service times occurs it is detrimental to patients waiting times and doctors idle time.

Hopp and Spearman (2001) distinguish random variation and controllable variation. Controllable variations occur as results of decisions and random variation occurs beyond control. For example, the time between customer demands are generally not controllable. An example of controllable variation is if several services are delivered to a patient, there will be variability service times. The effects of random variation are more subtle and therefore require more sophisticated tools to describe; therefore this study will focus on the random variation because these are unknown. Two types of random variability are process time variability and flow time variability. To be able to respond on variability it is important to know the causes behind the variability, in the manufacturing industry these causes are:

“Natural” variability Random outages Setups

Operator availability Recycle

Examples of natural variability are fluctuations in time due to differences in operators, machines, and material. In diagnostic clinics this can be fluctuations in time of the different cameras, fluctuations in time because of personnel who needs more time for an examination. Fluctuation in material can be caused at a diagnostic outpatient because radiopharmaceuticals are not ready. This research will focus on fluctuations in operators (personnel) and machines (cameras).

Random outages refer to unscheduled downtime. Unscheduled downtimes are for example breakdowns, operators being called away on emergencies, and running out of consumables. For the diagnostic clinic this can mean that the camera is defect, nuclear workers are away for other tasks, or failing of radiopharmaceuticals.

The third cause, i.e. setups, reflects the process changeovers. These are random variation when changes occur due to changes in the production process. Because diagnostic clinics have to deal with protocols, production processes cannot be suddenly changed.

Operator availability is the fourth cause Hopp and Spearman (2001) mention. When operators are not available the machine cannot run. This is partly true for a diagnostic clinic. When the camera is running it runs automatically, but the start and the end of the scan requires the availability of an operator.

The fifth cause, i.e. recycle, reflects to quality problems. This happens when a workstation performs a task and on the end checks whether the task was done correctly. If the task was not completed correctly, the task

(22)

22

will be repeated. This rework has major influences on capacity and also contributes on the variability of the examination times.

These five major causes for supply variability are based on the manufacturing industry, however as mentioned above these can be easily translated to the diagnostic clinic. According to Slack et al. (2004) the impact of uncertainty on estimating lead times leads some managers to use a probability curve to describe the estimation. Krajewski and Ritzman (2005) therefore it is necessary to set a service-level to determine the desired probability of not running out of stock. In healthcare, the manager may select a 95% service level. In other words, the probability is 95% that the estimated lead times will not exceed. By reducing, or eliminating supply variability the utilization level can be increased which in the end will results in shorter waiting times.

3.2.2.3 S

UB CONCLUSION

Which factors of supply influence the capacity planning and control? The answer on this question is supply variability and work design.

(23)

23 3.3 D

EPENDENT VARIABLES

3.3.1 C

APACITY PLANNING AND CONTROL

The purpose of planning and control is to run the operation processes effectively and efficiently and produce the service as required by the customer (Slack et al., 2004). The goal of an effective and efficient process is to increase the speed of added value processes (McNair, 1998). According to Grönroos (2006) value is created in services when customers (patients) use services. The goal of capacity planning and control is to reconcile demand and supply so added value processes can be maximized. According to Mango and Shapiro (2001) when the problems become clear it is usually possible to eliminate these problems by applying a better planning which reduces variability. Variability can exist on both sides; demand and supply. On the demand side, the important variability is the number of patients arriving at a certain time (Mango and Shapiro, 2001). On the supply side variability can exist because of poor deployment of resources or people (Mango and Shapiro, 2001). For example variability in preparation time; one nuclear worker prepares a patient in 15 minutes while another nuclear worker needs 30 minutes for the same task.

The reconciliation of demand and supply leads to better utilized resources (personnel and equipment) and minimized waiting times for internal (referring specialists) and external (patients) customers (Cayirli and Veral, 2003). According to Krajewski and Ritzman (2005) capacity planning is crucial for the long-term success of an organization.

3.3.1.1 R

ESOURCE ALLOCATION

Resources of a service are the facilitating goods, employee labor and capital (Fitzsimmons and Fitzsimmons, 2006). Resources have capacity, which means that a resource has the ability to generate production. Production can be measured in healthcare in terms of the amount of patients per unit of time (Vissers and Beech, 2005).

Matching fluctuating demand with available capacity is one of the most significant challenges for managers in any service industry (Jack et al., 2006). The allocation of limited capacity of resources among several customer types is a critical decision encountered by many service firms (Ayvaz and Huh, 2010). According to Naylor (1991) every healthcare system in every country, no matter how rich the country is, has to ration medical services because resources that have to be allocated are always limited. The inability of matching demand with available capacity can have serious consequences in the healthcare industry, like denying or limiting patients (Naylor, 1991). In other service industries this limit availability can have positive consequences because it creates an image of exclusivity. This exclusivity image makes it possible for an organization to raise fees. Health care can have serious consequences (life or death) and therefore hospitals are challenged to develop resource management strategies. These strategies enable them to provide consistently high quality services (Jack and Powers, 2009). This means that high quality services can be delivered during peaks and troughs.

(24)

24

FIGURE 3.5: ECONOMIC TRADE-OFF IN CAPACITY PLANNING (FITZSIMMONS AND FITZSIMMONS, 2006)

According to Vissers (1998) who studied resource allocation based on the patient-flow of inpatient resources, there are three main reasons for lost capacity. The first reason is when there is no balance between resource availability and demand high costs of service occur or high costs of waiting (figure 9). When there is to much capacity and low demand the cost of service will increase. In contrary when demand exceeds capacity the costs of waiting will increase.

The second reason refers to the timing of resource allocation which can lead to peaks and troughs (Vissers, 1998). Services have the characteristics ‘simultaneously’; this reflects the impossibility of holding inventory (Fitzsimmons and Fitzsimmons, 2005). Inventory is used in factories to decouple the stages of the process however decoupling in a service is achieved through customer waiting. Customers are waiting during the peaks and staff is waiting during the troughs.

The third reason of Vissers (1998) refers to the impact of capacity. When capacity of different resources, which are required simultaneously, are not in balance it will result in bottlenecks or under-utilization. Planning and control activities can balance demand and supply (Slack et al., 2004).

3.3.1.2 P

LANNING AND CONTROL ACTIVITIES

According to Slack et al. (2004) four planning and control activities can be distinguished: (1) loading, (2) sequencing, (3) scheduling, and (4) monitoring and control (figure 3.6).

FIGURE 3.6: PLANNING AND CONTROL ACTIVITIES (SLACK ET AL., 2004)

(25)

25

Loading

The first activity, i.e. loading, is the amount of work that is allocated to a work center (Slack et al., 2004). According to Schönsleben (2007) two types of loading can be distinguished: (1) finite loading and (2) infinite loading (figure 3.7 and 3.8).

As shown in figure 3.7 and 3.8, finite loading is an approach which allocates work to a work center up to a limit. In contrary infinite loading is an approach to loading work to work centers which does not limit accepting work but tries to cope with it. In healthcare infinite loading is not accepted because at the end of the day a diagnostic clinic cannot turn away the patients. Finite loading influences the utilization level because it is hard to achieve a utilization level near to 100% because diagnostic clinics also leaves slack in the planning for emergencies.

Sequencing

The second activity, i.e. sequencing, refers to decisions that have to be taken in what order things have to be done. Waiting lists are often the subject of debates (Wang, 2004). Some predefined rules are determined to give priorities (Slack et al., 2004; Gutiérrez et al., 2010). The priority of work may be determined by the physical constraints, customer priority, due date, LIFO (last in first out), FIFO (first in first out), LOT (longest operation time) or SOT (shortest operations time) (Slack et al., 2004). Patients face significant risks of complications or death when treatments are delayed (Wang, 2004). Therefore the process of prioritization is essential (Curtis et al., 2007) and can be used to maximize the welfare of all patients (Wang, 2004). According to Lim et al. (2005) a standardized protocol for prioritizing patients may be useful in helping to manage scarce resources in a fair and objective way.

The most well known priority in healthcare is customer priority or FIFO. This method contributes minimal to the waiting lists (Gutiérrez et al., 2010). In this case the administration will work through the examination requests, read them in sequence and schedule the patients in the same sequence.

Different prioritization systems can be applied within customer priority. According to Perris and Labib (2004) a distinction of prioritization can be made based on: (1) objective criteria (age of recipient or waiting time), or (2) subjective criteria (pain level, quality of life, etc.). The objective criteria are comparable with the critical criteria of the research of Curtis et al. (2007). This last model (subjective criteria) results into so called fuzzy logic. This model is a template for medical professionals that make clear that limitless additional criteria can be applied (Perris and Labib, 2004).

0 20 40 60 80 100 120 140 160 1 2 3 4 5 6 Lo a d i n h o u rs Work centers 0 20 40 60 80 100 120 140 1 2 3 4 5 6 Lo a d i n h o u rs Work centers Capacity Capacity

(26)

26

Priority may also be determined by due date. This means that work is sequenced according to when it is ‘due’ for delivery. This method improves the delivery reliability and also improves the average delivery speed (Slack et al., 2004). The con of this method is that it may not provide optimal productivity.

Scheduling

Scheduling is an excellent method when operations require a detailed timetable (Slack et al., 2004). Scheduling is a complex task in operations management because different types of resources need to be managed. Patient scheduling is the process of assigning patients to appointments. According to Cayrili and Veral (2003) two categories can be classified for appointment scheduling: static and dynamic. In the static case the decisions have to be made before starting the clinic session. In the dynamic case the schedule of arrivals continuously change over the day. This study is focused on static scheduling because radiopharmaceuticals have to be developed.

Outpatient clinics can be seen as queuing systems (Cayirli and Veral, 2003). The simplest case is when patients arrive at time and a single doctor serves the patient with stochastic processing times. When multiple doctors and multiple services are considered it becomes more complicated. Almost all literature is focused on single service systems where patients queue for a single service. In multi-service systems patient pass through different facilities such as registration, pre-examination, post-examination, x-ray, checkout, etc. For patients of the NMMI clinic this means they first pass through the preparation followed by the PET scan.

Schönsleben (2007) distinguished four types of scheduling: (1) forward scheduling, (2) backward scheduling, (3) central point scheduling, and (4) probable scheduling. To be able to schedule the planner needs to use lead-time scheduling techniques. According to Schönsleben (2007) this is a schedule that is developed by calculating the lead time. Lead time includes the operation time, interoperation time and administrative time in the manufacturing industry (figure 3.9).

FIGURE 3.9: LEAD TIME (SCHÖNSLEBEN, 2007)

Before an operation can run materials (patients) have to wait before being picked up (non-technical wait). Before running a technical wait time refers to the warm-up process of a work center. After the run time, technical wait time refers to the cool-down period and the non-technical wait time refers to the time before the material (patient) can be moved to leave the clinic.

The first type of scheduling, i.e. forward scheduling, refers to an order which is scheduled at the earliest start date and is completed at the earliest complete date.

The second type of scheduling, i.e. backward scheduling, refers to an order which is scheduled based on the latest completion date (due date) and starts at the latest start date.

The third type of scheduling, i.e. central point scheduling, refers to that the scheduled is determined based on the critical operation (bottleneck). This critical point determines the order schedule and therefore also the start and completion dates.

(27)

27

increased or decreased. Modifying slack can only be done by increasing or decreasing the non-technical interoperation times or the administrative times.

Monitoring and control

When a plan is created by loading, sequencing, and scheduling it is important to stick to the plan and ensure that the activities are indeed happening. ‘Drum, buffer, rope’ is a concept to decide where in a process control should occur (Slack et al., 2004). When processes do not have the same amount of workload onto each work center in most cases there is a bottleneck. The bottleneck is the drum, because it is the central point of the whole process. It is called the drum because this bottleneck sets the beat for the processes which are following. Bottlenecks are loaded all the time, therefore it is sensible to keep a buffer of inventory in front of it. Because services cannot hold inventory in terms of products, waiting customers are the buffer. The ‘rope’ refers to the communication between the bottleneck and the input to the process to make sure that activities before the bottleneck do not overproduce.

3.3.1.3 S

UB CONCLUSION

Based on the discussion in this paragraph about capacity planning and control we can conclude that the resource allocation and planning and control activities can influence the operational performance.

Vermeulen et al. (2009) stated that diagnostic resources in healthcare are central in the clinical pathways of patients. Therefore it is important to shorten or even eliminate waiting times for the NMMI service. Capacity planning can be seen as an economic trade-off. This trade-off ensures utilized resources and minimized waiting times (Cayirli and Veral, 2003). To be able to make this trade-off planning and control activities are necessary (Slack et al., 2004). Planning and control activities answer the questions; how much to do? When to do things? In what order to do things? Are activities going to plan?

To answer the question how much to do the organization has to choose between a finite loading and infinite loading approach. When to do things? This is an essential question because patients of the NMMI often face significant risks of complications or death when treatments are delayed (Wang, 2004). Because diagnostic resources are central in the clinical pathway, prioritization is useful to manage resources in a fair and objective way (Lim et al., 2005). The third question is in what order to do things? Patient scheduling is the process of assigning patients to appointments. Schönsleben (2007) distinguished four types of scheduling. To be able to schedule appointments it is important to calculate the lead time. In health care it is important to note that lead times can differ per patient or patient group. Modifying slack in schedules can be the solution. According to Schönsleben (2007) this can only be done by increasing or decreasing non-technical interoperation times or administrative times.

(28)

28 3.3.2 O

PERATIONAL PERFORMANCE

No standard performance measure is available for the healthcare field (Li and Benton, 1996). Everyone defines the performance of healthcare differently, based on their objectives, interests and interpretations. Patients for example find the quality of service important and a physician wants to achieve desirable clinical outcomes (Li and Benton, 1996).

From the operations management perspective the healthcare environment has recognized the need to reduce costs while the quality of healthcare has to increase (McLaughlin and Hays, 2007). Therefore hospitals have to manage their resources and processes more effectively and efficiently, which is the aim of operations management. The following four areas of operational performance can be distinguished (Hall, 2006):

Quality: conformance to specification

Flexibility: degree of coping with unpredictable situations

Speed: time between the moment of demand and the moment of delivery Dependability: degree of meeting the arrangements agreed

The book of Slack and Lewis (2002) mentioned a fifth area of performance which is costs. According to Slack et al. (2004) this area is affected by the other four performance areas (Slack et al., 2004).

The performance objective which is relevant for this study is speed. Be reminded that early diagnostics are essential for cancer patients for optimal treatment (Spiro et al., 2008). Therefore it is important that patients of a hospital service can be examined as quickly as possible.

Another performance criterion to be able to be efficient and effective is the degree of utilization. Capacity utilization is a measure how much output is actually used relative to the process capacity when fully busy (Fitzsimmons and Fitzsimmons, 2005, pp. 233).

Operationalizing these two performances can say whether an operation is effective and efficient.

3.3.2.1 S

PEED

When an examination for the NMMI clinic is delayed patients can face substantial risks of complication or even death (Wang, 2004). Therefore speed is an important performance objective. Delays often occur in health care. Almost all of us have waited for days or weeks to get an appointment with a physician for example. The only solution almost seems to be adding capacity (Van der Voort et al., 2010). Many health care practices are in state of disarray because the appointment books are overfilled. When an urgency patient has to be scheduled medical assistants, physicians and receptionist are mired in a telephone triage. According to Murray and Berwick (2003) a physician described the reception desk as a war zone with battles between patients and receptionists about appointments. Therefore it is hard to judge who is more stressed and dissatisfied; patients, receptionists, or physicians.

(29)

29

FIGURE 3.3: MILESTONE TIMES FOR THE DELIVERY OF A SERVICE (SLACK AND LEWIS, 2002)

According to Dickson et al. (2005) all service experiences require customers to wait at some point of the delivery process of a service. The definition of speed for this study is (Slack and Lewis, 2002):

This definition refers to the waiting time shown in figure 3.3. The faster the customer can have the service the greater the benefit they receive (Slack et al., 2004). In hospitals speed can mean life or death. According to Huang (1994) waiting times for outpatient services is the major reason that patients are dissatisfied.

However, little research is done to this type of waiting. The vast majority of literature reviewed the direct waiting times during a service session (Gupta and Denton, 2008). But in reality appointment systems influence direct waiting time and indirect waiting time. However, indirect waiting times are challenging to map accurately. First, it is challenging because patients struggle to find a suitable appointment time because of planned appointments with other physicians. The second complicating factor is that providers deal with access constraints because of the predefined sequence and mix of cases each day.

According to Vermeulen et al. (2009) short waiting times of local resources are important to achieve a high hospital-wide patient throughput. Because diagnostic resources are literally central in the clinical pathways of many patients it is immediately a bottleneck for health care processes in a hospital when long waiting times occur. According to Murray and Berwick (2003) the major barriers to perform efficiently are psychological; “the fear of change and the lack of confidence that existing resources can meet the demand for care” (pp. 1036). This can be empirically improved by the fact that the length of waiting lists remains steady at a given number of weeks or months. When supply is absolutely insufficient the waiting list will grow without limit.

Fast response to the customer is important but speed inside the operation is also important (Slack et al., 2004). Internal speed is for example caused by speedy decision making or speedy movement of materials. By