Faculty of
Economics and Business
Master Thesis
REALISATION OF TARGET WAITING
TIMES AT HOSPITAL CARE-PATHS
CORINE BONTE
University of Groningen
TABLE OF CONTENTS ... 2 ABSTRACT ... 4 1. INTRODUCTION ... 5 2. THEORETICAL FRAMEWORK ... 7 2.1INTRODUCTION ... 7 2.2DIAGNOSING FRAMEWORK ... 7 2.3CONCEPTUAL MODEL ... 8 3. METHODOLOGY ... 12 3.1INTRODUCTION ... 12 3.2DIAGNOSING FRAMEWORK ... 12
3.2.1 Determining relevant problem areas ... 12
3.2.2 Useful tools for determining relevant problem areas ... 13
3.2.3 Determining production control causes within problem areas ... 15
3.3DATA COLLECTION ... 16
4. RESULTS ... 18
4.1INTRODUCTION ... 18
4.2PROCESS DESCRIPTION PROSTATE CANCER CARE PATH ... 18
4.3RESULTS TARGET WAITING TIMES ... 19
4.3.1 Biopsy until diagnosis I ... 19
4.3.2 Diagnosis I until the first scan ... 20
4.3.3 MRI, CT and/or bone scan until Diagnosis II ... 21
4.3.4 Diagnosis II until operation or lymph node dissection. ... 23
4.3.5 Summary results ... 25
4.4DETERMINING PATIENT FLOW CONTROL CAUSES WITHIN PROBLEM AREAS ... 25
4.4.1 First two steps ... 25
4.4.2 Last two steps ... 27
4.4.3 Further investigated causes ... 30
4.4.4 Interview with planners ... 32
5. CONCLUSIONS ... 34 5.1INTRODUCTION ... 34 5.2RESEARCH QUESTION I ... 34 5.3RESEARCH QUESTION II ... 35 5.4METHODOLOGY ... 36 6. RECOMMENDATIONS ... 37 6.1INTRODUCTION ... 37
6.2RECOMMENDATIONS FOR THE STAKEHOLDERS ... 37
6.3RECOMMENDATIONS FOR FURTHER RESEARCH ... 37
REFERENCES ... 38
APPENDIX ... 39
APPENDIX I:DESCRIPTION OF PROSTATE CANCER CARE PATH ... 39
APPENDIX II:DATA FILTERING ... 41
APPENDIX III:DISTRIBUTION OF SCAN TYPES ... 43
Improving performance in terms of waiting time reliability is increasingly important for hospitals. Within a care-path targets are formulated for the waiting time of patients between every consecutive hospital visit. This research means to detect opportunities for improvement concerning the performance of a care-path in the North of the Netherlands with respect to waiting time reliability. Especially in the last steps of this care-path improvement
This thesis discusses a research project focusing on the reliability of patient waiting times, more specifically the waiting times of patients between consecutive hospital visits within a care path. Care paths are used in hospitals all over the world. A care path is a methodology for the mutual decision making and organization of care for a well-defined group of patients during a well-defined period.1 The aim of a care path is to enhance the quality of care across the continuum by improving risk-adjusted patient outcomes, promoting patient safety,
increasing patient satisfaction, and optimizing the use of resources (Vanhaecht et al, 2007). In a care path, standardized waiting times are formulated for the waiting times of patients
between every hospital visit. These standards are a result of meetings between patients’ action committee and personnel who are involved in this care path. For establishing the target times government standards are also taken into account.2 It is important that these standard waiting times are met.
There are several reasons for establishing and meeting standard waiting times. Within a care-path there are two categories of waiting time periods; patients in the first category are waiting for treatments, second category patients have to wait for diagnostic results. Nowadays in health care, patients want to have fast and convenient access to safe and effective treatments and diagnostics (Vissers & Beech, 2005). There is an increase of competition in the medical sector, which forces hospitals to monitor their costs, but also to remain reliable and
satisfactory to their patients (Vissers et al, 2007). For patients, their expectations are highest on medical and technical service (Vissers et al, 2001). Waiting times for treatments are a great influence on the success rate of the treatment. When examination is delayed and patients have to wait, patients can face substantial risks of complication or even death (Wang, 2004). Waiting times for diagnostics results can have influence on the well being of patients. Because cancer is a serious disease that can have disastrous consequences, patients can face much anxiety and fear during waiting times for results. Therefore it is important these waiting times are short and that patients are informed about the length of waiting time. Besides
optimizing waiting times, there are also reasons for optimizing utilization of resources. Due to the budget pressure, there is a strong focus on efficient use of resources, including equipment and personnel (de Vries et al, 1999). In combination with meeting target waiting times, a trade-off has to be made, because higher resource utilization leads to higher waiting times and higher variability in waiting times (Hopp & Speerman, 2008). For these reasons it is
important that hospitals have targets for the waiting times and tries to schedule process steps as effective as possible in terms of planning and utilization of resources to meet these target waiting times.
This research project provides the analysis of a real-life case. The first aim is to conclude to what extents real waiting times of patients comply with the criteria of standard waiting times. The second aim is to find causes for the percentage of patients not complying with target waiting times in order to find opportunities for improvements. In order to reach these goals, the framework developed by (Soepenberg et al, 2012) is used. Originally this framework is made for diagnosing the delivery reliability performance of make-to-order companies. So far, this framework is never used in a care-path setting. Because there are some similarities between make-to-order companies and a care-path with respect to the process, this project investigates whether this framework is useful in this setting.
1
The real-life case, which will be investigated, is the care path of ‘carcinoma of the prostate’ (hereinafter: prostate cancer care path or PCCP) at a hospital in the north of the Netherlands. The stakeholders for this case are personnel involved in this care-path consisting of: a nurse, planners, the organizational manager and urologists. Prostate cancer is de most common cancer that affects men in the Netherlands.3 It is seen as a serious disease and it has serious mental consequences. Therefore there is a great need for optimal medical care, guidance and information towards patients. This master thesis is written to inform and advise the
stakeholders in order to increase the reliability towards patients with respect to waiting times between hospital visits. First of all, the performance will be indicated by the percentage of patients who receive care within the prescribed target times. Based on these results, an investigation is done in order to find factors that negatively influence the waiting times for patients.
For the stakeholders of this care path the reliability in terms of meeting target waiting times is very important. Currently, a clear overview in meeting waiting targets is missing and the feeling exists that these target times are not met. With this research project I want to contribute to increase the level of reliability with respect to waiting times.
According to Soepenberg (2012), detecting improvement opportunities requires a structured diagnosis of the current performance. Therefore, the first step is to diagnose the current performance. The second step is to detect improvement opportunities, i.e. opportunities for waiting times of patients to comply with target waiting times. This leads to the following research questions:
RQ 1: To what extent do waiting times of patients in the care path meet the
prescribed target waiting times?
RQ 2: How can waiting times of patients comply with target waiting times?
Waiting times refer to the time in calendar days between every step in the care-path. The remainder of this paper is organized as follows: in Section 2, factors which influence reliability of waiting times is investigated which in the end result in a conceptual model. In section 3, the methodology is given by which the analysis is done. In section 4 the results of the first and a part of the second research question will be answered. In section 5 conclusions are given and in the last section recommendations are given for the stakeholders of the care-path as well as recommendations for further investigation.
3
2.1 Introduction
The objective of this research is to determine waiting times and to detect problem areas that reduce reliability with respect to waiting times. As mentioned before, to give an answer on the research questions, the framework developed by Soepenberg et al (2012) is applied.
Originally this framework is developed for diagnosing the delivery reliability performance of make-to-order companies. To apply operations management principles it is important to identify the assumptions underlying these principles and to discuss their applicability to hospital care processes (Vissers & Beech, 2005). Therefore, in this chapter similarities between make-to-order companies and care-paths are discussed to be able to assess the potential usefulness of the tool in a healthcare setting. Furthermore a theoretical foundation for causal relationships underlying reliability of waiting times will be discussed. This scientific knowledge also stems from make-to-order companies. In the end, the conceptual model is given.
2.2 Diagnosing framework
In make-to-order (MTO) companies, competition between companies creates high pressure on performance in terms of quality, efficiency and flexibility (Vissers & Beech, 2005). Therefore logistics-oriented manufacturing has often contributed to improvements in customer
performance like delivery time reliability as well as efficiency in terms of use of resources (Vissers & Beech, 2005). Health care is confronted with similar challenges as manufacturing (MTO) industry such as (de Vries et al, 1999):
- Need for efficient use of resources and reduction of costs.
- Increased pressure to improve quality of services by, amongst others, decreasing waiting times.
- Need to control the workload of nursing staff and other personnel. By comparing care-paths with MTO processes, there are some similarities:
- A process flow, which consist of several consecutive steps.
- Frequently there are targets established for the (maximum) throughput or waiting time for products or patients.
- In a MTO company, production start only when orders arrive, to reduce finished goods inventory. In addition because production start is delayed until ordering, the products can be a combination of customized and standardized components, in order for customers to purchase more exclusive orders. This holds also true within a care-path situation where handling starts only when patients arrive and specifications become clear during the process and can be different for each patient.
- In both cases there is little predictability because customers order specific products and patients have specific complaints, thus each product/patient has to undergo different actions. This makes planning rather difficult.
Besides similarities, there are some differences between the MTO and care-path situation. Figure 1 and 2 give a global representation of the MTO and care-path situation respectively. The main differences are as follows:
- MTO companies are focused on material flows, while the core process of the care-path is concerned with the flows of patients who need treatment (Vissers & Beech, 2005). A big difference in this is, patients have an opinion about their treatment which can influence the process.
- As for the MTO situation, important for customers is only the waiting time for the throughput of the whole process. It is not important how long each individual waiting time and production time takes. In the situation of the care-path, for patients,
especially the waiting times between each process step are important, because in most cases patients want to have their treatment or diagnosis as fast as possible.
- In MTO companies, the target throughput times can be different for each product. In general, for each ordered product the throughput time is established. Within the care-path situation target times are fixed, which means they apply to all patients. However, occasionally the actual waiting time is shorter when a doctor considers this necessary for a specific patient, for example when cancer is in a far advanced stage.
Operation 1 Operation 2 ... Operation n
Variable target throughput time
Figure 1: Process MTO-situation
Visit 1 Fixed target waiting time 1 Visit 2 Fixed target waiting time 2 ... Fixed target waiting time n Visit n
Figure 2: Process care-path situation
2.3 Conceptual model
Production control decisions should help to increase reliability performance with respect to throughput times (Soepenberg et al, 2012). Production control in health care organizations can be defined as: ‘The design, planning, implementation and control of coordination
mechanisms between patient flows and diagnostic & therapeutic activities in health service organizations to maximize output/throughput with available resources, taking into account different requirements for delivery flexibility and acceptable standards for delivery reliability and acceptable medical outcomes’ (Vissers & Beech, 2005, p27).
To achieve high reliability in waiting times, both the average and the variance of waiting times have to be controlled (Soepenberg et al, 2008). Some control decisions focus on shorter average waiting times while others focus on reducing the variance of waiting times
(Soepenberg et al, 2012). According to the design framework for production control
Gaalman, 1996). Thus, for a good performance on high reliability in waiting times, input and output, control decisions must be integrated. Input and output control decisions that could influence average and/or variance of waiting times are as follows (Soepenberg et al, 2008):
Input control decisions
- Target waiting time:
The decision on the amount of maximum waiting days for every step. - Patient acceptance:
The decision on the amount of patients starting the care-path. - Priority rules:
The decision on the sequence of patients.
Output control decision
- Adjusting capacity:
The (decision on the) availability of resources.
Input control decisions
Target waiting time
The target waiting times are established standards from higher management and
representatives of patient groups. This target waiting time has to be taken into account when patients enter each process step in the care-path. Therefore this control decision is an input control decision. This control decision cannot easily be influenced. However, there are no actual targets known for the percentage of patients that have to meet the standards. According to stakeholders approximately 80% would be sufficient. For the future it is important that this percentage is a fixed, measured number, since results can be quite different when this
percentage varies. For example, it is possible that 83% of the patients do not exceed target waiting times. Then it would be logical if the target for this percentage were 80% or 85%. In the first case, results are positive, in the second case they are not. However, in this research this control mechanism will not be addressed in the analysis due to the fact that for now this is not important.
Patient acceptance
Patient acceptance refers to the amount of (specific) patients during a certain period. The hospital’s highest priority is tending to people’s health, which makes treatment for its patients very important, consequently the hospital will not refuse a single patient, but will provide treatments for all. Therefore, this control mechanism has low preference.
Priority setting
time is reduced. When controlling the average of waiting times, a common used priority rule is the shortest process time rule (SPT) in which the patient with the shortest treatment time is treated first in order to reduce the average. However this rule is not preferable because first, the treatment time is difficult to estimate and second, this will increase the variance that could lead to very long waiting times for some patients and can increase risks for these patients which is absolutely not preferable. Overall, for the care-path the most preferable priority settings are FCFS and priority based on urgency.
Output control decisions
Adjusting capacity
This control mechanism is focusing on the availability of resources. Resources in the care-path are personnel, rooms, materials and equipment. Every resource has a capacity, which refers to the ability to treat patients (the amount of patients treated per unit of time (Vissers & Beech, 2005)). Capacity control has multiple levels. First, the allocation of resources to different departments, second the allocation of resources to different patient groups and last the allocation of resources to different patients within a patient group. Within the PCCP most resources are shared with other patient groups on the department of urology. Reasons to share resources can be costs, quality and the control of resource use. Sharing of resources will facilitate to reach the goal of high occupancy and productivity. There is a difference between time-shared resources and other shared resources. Time-shared resources are resources that are allocated to a specific patient group. This group can indicate preferred time periods for the resources, however its final allocation depends on the total of allocations that are made to all specialties. An example of a time-shared resource is an operating theatre. Other shared resources are generally available for all patient groups and do not have a special allocation arrangement. The allocation of resources is a difficult but important issue for effective and efficient hospital operations management (Vissers & Beech, 2005). Most hospitals do not have a good method to allocate their resources in an effective and efficient way. Especially the operating rooms are important resources in the hospital and at the same time difficult to allocate. It is difficult because the work performed is very labour-intensive and it involves very expensive materials and equipment. Therefore in many cases the operating rooms are the bottleneck resources. A bottleneck resource is the resource that is most scarce and therefore determines the overall volume of production.
The third level of capacity control is the allocation of resources to different patients within a patient group. Treatments can be characterised by the time of treatment (duration) and by the number and types of resources required (workload). These characteristics affect resource requirements and thus the way treatments can be planned and controlled. In order to control the average and variance of waiting times, it is important to dedicate capacity to those
resources where patients are congested and to meet the peak requirements of specific patients respectively (Soepenberg et al, 2012).
Uncontrollable circumstances
The section above has concentrated on the influence of control decisions on waiting time reliability performance. This theory is summarized in a conceptual model given in figure 3.
Coordination between input & output control decisions Decisions on adjusting capacity Reliability of meeting target waiting times Decisions on target waiting times Input control decisions Decisions on patient acceptance Influence Deciosions on priority setting Output control decisions part of part of part of part of Uncontrollable circumstances Influence
Figure 3: Conceptual Model
3.1 Introduction
Identifying improvement opportunities requires a structured diagnosis of the current performance (Soepenberg et al, 2012). In this research, the framework developed by
Soepenberg et al (2012) for diagnosing the delivery reliability performance is applied. In the following paragraph this diagnosing framework is presented and its applicability in the care-path setting is discussed. The third paragraph discusses the data collection which is required for applying this methodology.
3.2 Diagnosing framework
The diagnosing framework developed by Soepenberg et al (2012) consists of two parts. In the first part, the relevant problem areas are determined by executing four steps - this part is described in the first sub-paragraph. The second sub-paragraph explains the tools that can be used by executing the fourth step of determining relevant problem areas in the MTO - as well as the care-path situation. In the second part of the diagnosing framework, causes of bad performance due to production control decisions are determined; these are discussed in the third sub-paragraph.
3.2.1 Determining relevant problem areas
In this sub-paragraph the steps of determining relevant problem areas are provided. For the first three steps it is questioned whether causes of bad performance must be sought for in the average or variance of lateness. Based on this result, the diagnosis should be average-oriented or variance-oriented. In the fourth step it is determined in which direction analysis should continue.
Step 1
In the first step of the first part of the diagnosing framework the percentage of orders or patients delivered or helped too late in a certain period is determined. The promised delivery time is compared to the real delivery time for each product/patient. Based on the lateness values of all orders or patients, a distribution of lateness can be constructed (see figure 4) (Soepenberg et al, 2012). Lateness is defined as the conformity of a schedule to a given due date (Baker, 1974). Lateness can be distinguished by positive (a patient is served later than the target time prescribed) and negative lateness (a patient is served earlier than prescribed). In the care-path situation for every step for which target waiting times are available, the target waiting time is subtracted from the real waiting times of patients. On the basis of this data a distribution of lateness can be constructed and the percentage of patients complying with the target waiting times can be established. This step will give the answer on the first research question.
Step 2
When there is a high average of lateness, it is checked whether a limited number of patients is responsible for this. In the care-path, it could be that in particular cases patients who delay the appointment themselves cause the high average. When there is a high variance of lateness two situations are optional. 1, some specific patients have a high internal variance of lateness, which influences the total variance of lateness. In this case further analysis should be executed for this specific patient group. 2, the internal variance of different patient groups is limited and the variance of lateness is mainly caused by the differences in averages of the patient groups. For example, when the variance of the total waiting times of patients is fairly high but the variance of waiting times of emergencies and normal patients is limited. In that case there is a large difference in average waiting times between these two groups. In this case, analysis should proceed by focusing on the group of patients with high average.
Step 3
According to Land, (2004) in a manufacturing environment, achieving high delivery reliability is a combination of controlling average throughput times and controlling the
progress of individual orders. Useful tools to control average throughput times and controlling the progress of individual orders are the throughput diagram and the order progress diagram, developed by Soepenberg et al (2008). In this step the throughput diagram is used to measure the average lateness and the order progress diagram is used to get insights into the variance of lateness. Lateness can be defined as the conformity of a schedule to a given due date (Baker, 1974). The throughput diagram and the order progress diagram are further explained in chapter 3.2.2.
Step 4
Depending on the decisions made in the steps described above, the diagnosis will be either average-oriented or variance-oriented. The fourth step is executed to determine whether the diagnosis should focus on the process of promising delivery time and/or the realisation process. In a care-path case, the delivery time promising is the waiting time promising. This waiting time promising process cannot easy be influenced which is earlier described in the theoretical framework (see chapter 2). Therefore, when target times are not met, the focus should be on the realisation process instead of the waiting time promising process.
3.2.2 Useful tools for determining relevant problem areas
For a complete diagnosis of waiting time reliability, the analysis of the average lateness as well as the analysis of the variance of lateness should be performed. Throughput diagrams can support analysis of the average lateness. The order progress diagram is an instrument that can be used to analyze the variance of lateness. In this sub-paragraph these tools are explained. First the tools are explained for the MTO-situation, thereafter for the care-path situation.
Throughput diagram
MTO-situation
work waiting). Only when the FCFS (first in first out) method is used, the horizontal distance indicates the throughput or lead times of products.
Figure 5: Throughput diagram (MTO-company)
Care-path situation
For the care path situation a throughput diagram can be made for the waiting time in-between two consecutive visits of the care-path. The horizontal axis represents cumulative time in calendar daysand the vertical axis represents waiting measured in cumulative amount of patients. The input curve represents the cumulative amount of patients who finished the first visit. The output curve represents the cumulative amount of patients who finished the second visit. Let’s assume we draw a throughput diagram for the waiting time between visit 1 and visit 2 for a group of patients. The cumulative input is drawn based on all dates given of visit 1. The cumulative output is drawn based on all dates of visit 2. The output curve increases on a certain date by the number of patients who went for their second visit on that specific day. The vertical distance between the two curves represents the amount of patients waiting for their second visit on that specific moment. The horizontal distance represents the real waiting times per patient, only when the FCFS discipline is applied. The FCFS method means patients will be served at their second visit in the same order as they were served at their first visit. A throughput diagram gives only the performance in terms of the average lateness, but does not give the variance of lateness (Soepenberg et al, 2008). The variance of lateness can be analysed by an order progress diagram that is discussed below. An analysis based on throughput diagrams can indicate in which specific periods delays of patients’ waiting time are caused by congestion of patient flows (Soepenberg et al, 2012).
Order progress diagram
MTO-situation
Figure 6: Scatter diagram
An example of the order progress diagram is given in figure 7. The progress of a single order is represented by one curve. The distance between the left dot and the horizontal axis
represents the estimated lateness at the moment the order arrived at the first step. This is calculated by the promised throughput time minus the average throughput time. The right end dot represents the real lateness at the end of the process. The upward or downward direction of each line segment represents respectively an increase or decrease of estimated lateness of the order in that particular stage. The order progress diagram is used to indicate the causes why some orders are delivered too late and why some orders are delivered too early (Soepenberg et al, 2008).
Figure 7: Order progress diagram
Care-path situation
For every step for which target waiting times are available, a scatter diagram can be made. These scatter diagrams show when patients were helped too late in this stage of the care-path. Furthermore, also an order progress diagram can be made for each waiting moment of the care-path. Since the promised/target waiting time is the same for every patient, the left dot at the beginning of the care-path starts at the same estimated lateness for every patient. This is because it is calculated by the target waiting time minus the average waiting time, which is the same for each patient. Therefore, in this case, the right end-dot of every step provides all information necessary. As a result the choice is made to only use order scatter diagrams as shown in figure 6.
3.2.3 Determining production control causes within problem areas
In the first step of the diagnosing framework, delivery reliability is measured and relevant problem areas are identified. As mentioned in step four of the diagnosing framework, the focus is placed on the realization process of target waiting times. In the second step
an average-oriented diagnosis within the realization process can be executed; second a variance-oriented diagnosis can be performed. Based on the throughput diagrams, scatter diagrams and/or order progress diagrams, the diagnosis can be executed. For example when the throughput diagram looks similar to figure 8, the (average) waiting times for patients increases when real time increases. The cause for this should be investigated. It can be possible there is insufficient capacity. For detecting causes of negative influences, both general problem-solving knowledge and domain knowledge is required (Wagner, 1993).
Figure 8: Throughput diagram
In order to clearly define and limit the research, the focus for analysing is on those patients who were helped in the operating room. At this moment stakeholders experience high uncertainty in complying with target waiting times for patients before being treated in the operating room. According to theory, input and output coordination at the operating room is challenging. The operating room and the specialists are shared resources, which means several patient groups use these resources. Specialists can also be used within other care-paths. Therefore, specialist and the operating room are not always available for this care-path. This makes coordination more difficult. In general, the operating theatres and staff are the most important resources of the hospital to produce its service (de Vries et al. 1999). According to Vissers & Beech (2005), this is because the work performed in the operating room is very labor-intensive and involves expensive equipment and materials. For these reasons the focus is on analyzing patients who were treated in the operating rooms.
3.3 Data collection
For the diagnosis, quantitative data is needed on both target waiting times and patient’s progress. In the hospital, several years ago they started a database to registrate all processes for patients. The main goal of the registration of this data is for finance transactions. Because of financial purposes, every medical action is registered, which results in tens of thousands of data points. This makes it difficult to use for the purpose of this research. After several meetings with the stakeholders, this data is made more understandable and it is filtered to be able to execute this research. The filtering process of the data is more accurately described in appendix II. The data is gathered for a period of 2 years, for 2010 and 2011. For the first step of the diagnosis framework, all data that meet requirements is used. For the second step, in general only data of patients who had an operation or lymph node dissection are used, in order to restrain the research as mentioned before and to see what happens to these patients. After analyzing, results of analysis are presented to the stakeholders to verify the outcomes and eventually discuss further clarification on those factors that could not be measured
4.1 Introduction
In this chapter results of the analysis are reported. In the second paragraph the care path will be described. In the third paragraph the parts for which target times are formulated are shown and the percentages of patients who comply with target times are given, which will answer the first research question. In the last paragraph the second research question will be addressed in which some examined causes for not complying with the target times are clarified.
4.2 Process description prostate cancer care path
Immediately when the possibility of prostate cancer arises and a patient is sent to the hospital, the patient proceeds to an intense care-path with several treatments and diagnostic
examinations. This care-path is called the ‘prostate cancer care path’ (PCCP).
A general representation of PCCP including target waiting times in-between several hospital visits is given in figure 9 below. These target waiting times are shown in red.
Presumption general practitioner Diagnosis I Results 10 days Metastatised cancer No cancer Not metastatised cancer End Lymph node dissection cancer Diagnosis II Results Operation Scans 7 days Biopsy 10 days 14 days 14 days
Figure 9: Simple version PCCP
The care-path is part of the urology department. PCCP covers about 17% of total patients at the urology department. The care-path starts with a referral to the hospital as soon as there is suspicion of prostate cancer by a general practitioner. At the hospital, the suspected prostate cancer is typically confirmed by taking a biopsy of the prostate, this is a medical test
involving sampling of cells for examination under a microscope.4 Based on this test, it is
4
diagnosed whether a patient has cancer. The result of this test is discussed with the patient during the visit of diagnosis I.
When cancer is confirmed, further tests are conducted in order to investigate whether cancer is metastasized (spreading of cancer to other non-adjacent organs or parts5). This test is executed by an MRI scan, CT scan or bone scan. This results in diagnosis II - in which results are discussed with the patient. When cancer is metastasized, the patient receives radiation therapy. Before the radiation therapy has started, a lymph node dissection is executed. In case the cancer is not metastasized, the patient undergoes surgery at this hospital or a different hospital. In this surgery all the cancer that can be removed will be removed.
See Appendix I for an extensive description of the care path.
4.3 Results target waiting times
In this paragraph research question one will be answered: ‘To what extent do waiting times of
patients in the care path meet the prescribed target waiting times?’ In order to answer this
question Excel files received, (which are further described in Appendix II) are used for analyzing. Four steps will be analyzed. Originally stakeholders would measure one additional step, which was the step of patient visiting general practitioner until the first visit of the care-path. Due to the fact that there is no information on time of appointments with the general practitioner, the waiting time between this visit until the first hospital visit cannot be measured. Therefore, this step is not taken into account for this research.
The relevant steps and its target waiting times are discussed during meetings with the stakeholders and are as written below:
1. Biopsy until diagnosis I; 10 days. 2. Diagnosis I until the first scan; 10 days. 3. Last scan until diagnosis II; 7 days.
4. Diagnosis II until operation or lymph node dissection; both 14 days.
In order to produce results, some choices have been made in the filtering of the received data. In appendix II, the procedure and clarification of filtering is presented. In the remainder of this paragraph, results are shown per step and a summary of results is provided.
4.3.1 Biopsy until diagnosis I
The target waiting time between biopsy and the results review in diagnosis I, is ten days. There were 419 biopsies registered during 2010/2011. After the filtering process 352 are used
for analysis. In figure 10 results are shown.
Figure 10: Distribution of lateness of waiting times between biopsy and diagnosis I Number of patients: 352
Target time (days): 10
Average waiting time (days): 8,73 Standard deviation: 2,68
Percentage of meeting target time: 76%
Waiting times are calculated between biopsy and diagnosis I for every patient that took both these steps. Following, actual waiting times are compared with target waiting times. In the figure the distribution of lateness is given in a frequency graph with columns. Below the figure, the number of patients, the target time the average waiting time and the standard deviation are given. The standard deviation shows how much variation exist from the average. Finally, also the percentage of patients who reaches the target waiting time is provided. The horizontal axis represents the waiting time in calendar days compared with target waiting times. On the vertical axis the amount of patients is represented. This means, 0 on the horizontal axis represents the amount of patients waited according to the target time. All columns higher than 0 represent the amount of patients who are helped too late. These
patients are represented by the red bars. The figures in the following sub-paragraphs have the same structure.
As can be observed from figure 10, the largest number of patients (see the largest bars), are two or three days too early. However even a reasonable number of patients (around fall within the area of one till three days too late.
4.3.2 Diagnosis I until the first scan
After the first result review, a number of patients have to undergo the scan. Several patients have to undergo a second scan, when the first scan does not provides sufficient result. In this analysis, only the time between diagnoses I and the first scan is analyzed. 225 patients were registered with a scan during 2010/2011. From this 225 there are only 145 also registered with a date for diagnosis I. After executing needful filtering 87 patients remain for the analysis. The distribution of lateness for this step is shown in figure 11.
0 10 20 30 40 50 60 70 80 90 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 A m ou n t of p ati e n ts
Figure 11: Distribution of lateness of waiting times between diagnosis I and scan 1 Number of patients: 87
Target time (days): 10
Average waiting time (days): 6,84 Standard deviation: 2,91
Percentage of meeting target time: 89%
As can be observed from figure 11 most of the patients are treated 4 days too early. The average waiting times is well below the target waiting time, which is also indicated below the figure. One remarkable thing is, no patient waited exactly according to the target waiting time.
4.3.3 MRI, CT and/or bone scan until Diagnosis II
For patients who have to go through multiple scans (in general not more than two scans), there are 2 possible scenarios:
Scan I Conversation Scan II Conversation
In this situation there is a conversation after scan I as well as scan II. In this case the second conversation is the official diagnosis II.
Scan I Scan II Conversation
In this situation there is only one conversation after scan II. This conversation is diagnosis II. For the analysis of the step from scan until diagnosis II, two scenarios are measured. These are discussed during meetings with stakeholders. The scenarios that are measured are as follows:
1. The last scan until diagnosis II.
2. The first scan until the first follow up appointment for patients with two scans.
The first scenario is measured in order to find out if target times are met. The target time is seven days. The second scenario is measured in order to see how long patients with two scans have to wait on the results of their first scan. For 225 patients, 309 scans are registered.
0 5 10 15 20 25 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 A m ou n t of p ati e n ts
There are three types of scans, CT-scans, MRI-scans and bone-scans. The appearance of each different scan during 2010/2011 is given in Appendix III.
Last scan until diagnosis I
In the following figure, for the above mentioned 166 patients, an analysis is performed. Results are shown in figure 12.
Figure 12: Distribution of lateness of waiting times between last scan and diagnosis II for all patients Number of patients: 166
Target time (days): 7
Average of waiting time (days): 8,23 Standard deviation: 3,32
Percentage of meeting target time: 47%
At this point, a remark must be made. As can be seen in the prostate cancer care path in appendix I, the target time between diagnosis I and diagnosis II is seventeen days, however with a maximum of twenty days. It is not obvious how to divide the three days between the two steps, so far this maximum is not taken into account. Noteworthy is that when these three days are added to the target waiting time of this step, the percentage of meeting the target time will increase to 81%, which is a large difference of 47%.
Looking at figure 12, we can conclude most of the patients waited exactly according to the target waiting time. In addition, the figure resembles a clock model, as we expect from a normal distribution. Therefore it would be that around 50% is too late, which is true.
The first scan until the first following appointment for patients with 2 scans
For patients who received two scans, the distribution of lateness is made for the time between their first scan and getting their results. This is shown in figure 13. 46 patients are analyzed.
0 5 10 15 20 25 30 35 40 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 A m ou n t of p ati e n ts
Figure 13: Distribution of lateness of waiting times between first scan and appointment for patients with 2 scans Number of patients: 46
Target time (days): 7
Average of waiting time (days): 10,04 Standard deviation: 3,85
Percentage of meeting target time: 33%
In this case, when target time was 10 days, 54% of the patients get their result within the target time.
As follows from figure 13, there is a big variance in patient’s waiting time. This is confirmed by the high value of the standard deviation. Note that the number of patients has been reduced somewhat in relation to the previous figure. The length differences of bars within the graphs are related to the total number of patients analyzed.
4.3.4 Diagnosis II until operation or lymph node dissection.
The final step is to analyze patients‘ waiting time between diagnosis II and the operation or lymph node dissection. Starting with 65 patients, after filtering 27 patients could be analyzed for the operation and eight patients for the lymph node dissection. The target waiting times for both cases are fourteen days. Since there are only a few patients registered with varied results, the distribution of lateness does not show a clear pattern. Figure 14 shows results for the patients who had to undergo surgery, figure 15 shows results for lymph node dissection patients. Operation 0 1 2 3 4 5 6 7 8 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 A m u on t of p ati e n ts
Figure 14: Distribution of lateness of waiting times between diagnosis II and operation Number of patients: 27
Target time (days): 14
Average waiting time (days): 24,81 Standard deviation: 9,52
Percentage of meeting target time: 15%
As can be observed from figure 14, there is much variation in waiting times. Furthermore there are too few patients and therefore it is difficult to make some more conclusions based upon this figure.
Lymph node dissection
Figure 15: Distribution of lateness of waiting times between diagnosis II and lymph node dissection
0 1 2 3 4 5 6 -13 -10 -7 -4 -1 2 5 8 11 14 17 20 23 26 A m ou n t of p ati e n ts
Lateness in waiting times in days
0 0,2 0,4 0,6 0,8 1 1,2 -13 -10 -7 -4 -1 2 5 8 11 14 17 20 23 26 A m ou n t of p ati e n ts
Number of patients: 8 Target time (days): 14
Average waiting time (days): 28,25 Standard deviation 6,54
Percentage of meeting target time: 0%
For this step, the same holds as for the analysis of the figure 14. There is much variation in waiting times for these patients. Furthermore, no real conclusions can be made.
4.3.5 Summary results
In table 1, a summary of the results of the first analysis is given.
Table 1: Summary results
For the step of the last scan until diagnosis II data is provided for the situation on seven days target waiting time as well as ten days target waiting time. The results for the ten days target are shown in parentheses. For the remainder of the research, a target waiting time of seven days is assumed.
A remarkable result shown in table one is the reliability performance in waiting time for the last two steps (from the last scan until operation or lymph node dissection) is very low.
4.4 Determining patient flow control causes within problem areas
For the second research question: ‘How can waiting times of patients comply to target waiting
times?’, the patient flow control causes within problem areas are investigated. As mentioned
in the theoretical framework, the focus will be on two control decisions. These two control decisions are priority rules (an input control decision) and adjusting of capacity (an output control decision). The main problem areas detected in the previous section are the final two steps. In this paragraph therefore the emphasis will be on these steps. In the first
sub-paragraph, control causes for the first two steps is attempted to be established and in the second sub-paragraph control causes for the last two steps. In the third sub-paragraph further investigated control causes for the whole process are attempted to be established. The last sub-paragraph discusses some remarkable things which were discussed with two planners during an interview which is processed. The first planner is from the radiology department, another of the urology department. The planner of the radiology department is responsible for the planning of amongst others, the scans. The planner of the urology department is
responsible for the planning of all appointments taking place on this department. For both departments these include all patients and are not limited to prostate cancer patients. The entire interview results are provided in appendix VII.
4.4.1 First two steps
long, the FCFS discipline could be applied, which focuses on reducing the variance of lateness. The use of this priority rule can have only a positive effect on the percentage of patients complying to target times, if average waiting time is below the target waiting time, which is the case. In this sub-paragraph the results of applying the FCFS rule are described. There is tried to find out if applying this rule would lead to an increasing percentage of patients complying with the target waiting time.
Biopsy until Diagnosis I
In figure 16, both the normal situation (as given in figure 10) and the situation for FCFS are shown for biopsy until diagnosis I. The dark blue and dark red bars represent the distribution of the normal situation as it was currently (see figure 10). The light blue and light red bars represent the distribution if the FCFS method would have been applied.
Figure 16: Distribution of lateness of waiting times between biopsy and diagnosis I with FCFS rule Number of patients: 352 Number of patients: 352
Target time (days): 10 Target time (days): 10
Average of waiting time (days): 8,73 Average of waiting time (days): 8,80 Standard deviation:2,68 Standard deviation:2,06
Percentage of meeting target time: 76% Percentage of meeting target time:77%
By adapting the priority rule FCFS; there is no real improvement in percentages. The percentage of meeting the target time is increased by only one percent. What showed to be remarkable - by adapting FCFS a great number of extra patients are late. (See the light red bars belonging to the lateness of one or two days.) This could mean there were periods where some patients received a priority that led to diagnosis on time. While at the same time other patients had to wait too long. This made the average lateness in that period too high. Through application of the FCFS rule, average waiting times will still be too high. In order to
determine in which periods this high average occurred, a throughput diagram is made which is shown in appendix IV. This shows that moments when the average waiting time exceeds the target waiting time are short and it happens at random moments. The solution for this in general, is making good output control decisions based on capacity. However, since the focus is on the last two steps, this will not be further investigated.
Diagnosis I until the first scan
In figure 17 the FCFS method is applied for the second step.
0 10 20 30 40 50 60 70 80 90 100 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 A m ou n t of p ati e n ts
Figure 17: Distribution of lateness of waiting times between diagnosis I and scan 1 with FCFS rule Number of patients: 87 Number of patients: 87
Target time (days): 10 Target time (days): 10
Average of waiting time (days): 6,84 Average of waiting time (days): 6,84 Standard deviation:2,91 Standard deviation:2,70
Percentage of meeting target time: 89% Percentage of meeting target time:90%
In addition in this step there is no real improvement in percentage of complying patients. Remarkable is, the most right light red bar, is increased. This is because now some outliers which originally were out of scope, re-entered the feasible region of the lateness distribution diagram. For this step there is also a throughput diagram drawn, which is shown in appendix IV.
Since no solution can be found by adapting the FCFS method, other opportunities for improvements need to be established. Of course the question is if waiting times already are good enough. As mentioned earlier, the overall feeling among stakeholders is that 80% would be sufficient, when this would be the case, then this track performs right.
4.4.2 Last two steps
Patients who followed the last two steps had to wait, on average, too long. Therefore, the FCFS discipline will have not a positive effect. Currently the applied priority rule is on basis of urgency of patients. Because other priority rules are not preferable, the focus will be on the availability of capacity.
In order to analyze potential causes for delays in the last two steps, a more in-depth analysis is conducted on the 40 patients included in this step. These 40 patients consist of 31 patients who went for an operation, 9 of them went for a lymph node dissection. After showing the first results as given in paragraph 4.3 to the stakeholders, two options for not complying to target times were provided:
0 5 10 15 20 25 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 A m ou n t of p ati e n ts
1. Patients occasionally want extra appointments before they decide to take an operation or lymph node dissection. This is one of the uncontrollable circumstances.
2. In periods of holidays there is insufficient capacity, which influence the lateness of patients.
Both these options are investigated. In addition other possible causes for positive lateness are investigated.
Option 1: extra appointments before operation or lymph node dissection
The first issue investigated, is how many patients did receive extra appointments before the operation or lymph node dissection. Eight out of forty patients (20%) arranged multiple appointments after diagnosis II. In table 2 these patients are shown with dates of the last scan, all appointments afterwards and the date for their treatment.
Table 2: patients who received extra appointments
For these eight patients, the average waiting time between diagnosis II and operation/lymph node dissection is 52 days, which is much higher than 27 days, the average of total patients which is shown in table 4. In table 3 the new lateness is shown for these eight patients when the last appointment is taken as the date for diagnosis II, compared to their real lateness.
Table 3: lateness patients who received extra appointments
When taking the last appointment as diagnosis II, the variance and average of lateness for operation or lymph node dissection is reduced. Still most of the patients are treated too late; only the green highlighted patient is currently treated on time. The influence on the total percentage of complying with the target time therefore is minimal. Results of influences on performance is shown in appendix V.
For the following results, these eight patients are not taken into account for analysis. This, because these patients may influence outcomes.
Number Patient Date last scan Diagnosis II 2nd appointment 3th appointment 4th appointment Treatment date
1 lymph node dissection 26-5-2010 1-6-2010 11-6-2010 - - 28-6-2010 2 lymph node dissection 8-4-2011 13-4-2011 27-4-2011 - - 13-5-2011 3 lymph node dissection 8-6-2011 9-6-2011 4-8-2011 16-9-2011 28-9-2011 4-10-2011 4 operation 25-6-2010 29-6-2010 2-7-2010 - - 23-8-2010 5 operation 12-5-2011 19-5-2011 27-5-2011 - - 6-7-2011 6 operation 21-7-2011 29-7-2011 20-9-2011 - - 7-11-2011 7 operation 13-10-2011 24-10-2011 26-10-2011 2-11-2011 - 30-11-2011 8 operation 6-10-2011 17-10-2011 9-11-2011 - - 14-12-2011
Number Official lateness Lateness new
Option 2: holidays
Last scan until diagnosis II
The other suggestion is the influence of holidays. Within the holidays the capacity of scans and operating room is decreased by 50%. For detecting dependency between the lateness and holidays, scatter and throughput diagrams are created. Figure 18 shows the scatter diagram from the last scan until diagnosis II for 32 patients at the moment they were treated at diagnosis II. When during holidays patients had to wait longer than normal, on the scatter diagram the dots (which represents patients) around holidays have to show a higher positive than the other dots.
Figure 18: Scatter diagram last scan until diagnosis II
As can be seen above, there is no special trend in time. Almost all dots are randomly located. During the holiday seasons there is no increased lateness. So, it can be stated that the
dependency between lateness and holiday periods is low for the step last scan until diagnosis II. Still 47% of all patients (see chapter 4.3.4) are too late. Another reason for this could be specialists who have insufficient time for analysing scan-results. This would be an option for further investigation. Since there are some remarkable outliers, these outliers are summarized in table 5.
Table 4: Outliers
Looking further into the details for the first three patients, it appears these patients did not follow the official way: they received their operation very fast after the last scan, without having a diagnosis II visit. This means the diagnosis II date is not the official diagnosis II date. The reason for this error lays in the fact that diagnosis I and diagnosis II are not officially registrated (see appendix II). From a file with all registered conversations, every
-10 0 10 20 30 40 50 22-1-2010 2-5-2010 10-8-2010 18-11-2010 26-2-2011 6-6-2011 14-9-2011 23-12-2011 Es ti mate d la ten es s in da ys Dates
Number Patient Biopsy Diagnosis I 1st scan Last scan Diagnosis II Operation
1 operation 1-4-2011 11-4-2011 5-4-2011 18-4-2011 19-5-2011 26-4-2011
2 operation 24-2-2010 4-3-2010 12-3-2010 - 8-4-2010 23-3-2010
3 operation 21-7-2010 29-7-2010 3-8-2010 - 24-8-2010 10-8-2010
4 operation 30-8-2011 20-9-2011 23-9-2011 11-10-2011 1-11-2011 1-12-2011
diagnosis II respectively. In some cases no official diagnosis took place. This is possible because sometimes patients receive their results by phone since these patients needed to be treated immediately. For the other two patients no obvious reason can be determined for the high lateness.
Diagnosis II until operation or lymph node dissection
Figure 19 shows the lateness between diagnosis II and the treatment, for the same patients, except the three patients with an error in registration of diagnosis II as detected before.
Figure 19: Scatter diagram diagnosis II until operation or lymph node dissection
In addition, this figure does not show any relationship between holiday periods and lateness. Because one of the planners commented that the capacity during holidays is reduced by 50%, you would expect more lateness around holidays. Possibly prostate cancer patients, do not suffer from this reduction of capacity, because they are getting priority compared to other patients. But still the patients are treated too late.
4.4.3 Further investigated causes
Besides analysis on each step separately, also analysis is conducted for the relationship amongst these steps. In figure 20 a scatter diagram is shown in which the lateness is given for patients for the whole process from biopsy until operation/lymph node dissection. The
Figure 20: Scatter diagram biopsy until operation or lymph node dissection
31% of the patients comply to the total target waiting times of 41 days. As can be seen in figure 20 the variance on lateness is very high. The two patients represented by the red dots are further investigated, due to their extreme values .The lower dot represents a patient who is much too early (29 days) and the higher dot represents a patient who is much too late (51 days). Both patients required an operation. Their treatment history is summarized in table 6.
Table 5: history patients of two red dots in figure 19
The patient who is too early is the same patient as the patient number one in table 5. As previously noted, the date for diagnosis II is not correct. This patient received a biopsy on 1-4-2011 and is treated at each step promptly. The patient, who is too late, was extreme delayed only at the end of the process. Until diagnosis II he received his treatments and diagnosis reasonably on time, but the time of operation was very much delayed. Causes for this can be the patient delayed it himself because of, for example holidays. It is not possible his lateness is related to capacity insufficiency because in the period after his diagnosis II, at least four patients were helped for who the lateness is much lower, occasionally even negative. Therefore it has more to do with priority.
Another cause of variances in lateness could be that sometimes patients get priority at the start, when there is a considerable suspicion of a bad cancer condition for the patient.
Therefore the relationship between the lateness from biopsy until scans and the lateness from the last scan until operation or lymph node dissection, which is given in figure 21. The black line represents a linear trend line. A linear trend line represents the relationship between a scalar dependent variable y and another explanatory variable x. In this case the first variable is the lateness of from byopsy until first scan and the second variable is the lateness of last scan until operation or lymph node dissection. The correlation value r = 0,22. A value of -1 means there is a strong negative correlation. A value of 1 means there is a strong positive corelation. The coefficient of determination is r2 = 0,222 = 0,05, which means that 5% of the total
-40 -30 -20 -10 0 10 20 30 40 50 60 22-1-2010 2-5-2010 10-8-2010 18-11-2010 26-2-2011 6-6-2011 14-9-2011 23-12-2011 1-4-2012
Patient Byopsy time between Diagnosis I time between scan 1 scan 2 time between Diagnosis II time between Operation
Too early 1-4-2011 10 11-4-2011 -6 5-4-2011 18-4-2011 31 19-5-2011 -23 26-4-2011
Too late 10-3-2011 11 21-3-2011 4 25-3-2011 25-3-2011 6 31-3-2011 71 10-6-2011
variation in the lateness last scan until treatment can be explained by the linear relationship between the lateness last scan until treatment and the lateness byopsy until first scan. The other 95% remains unexplained. Therefore, the relationship between these two lateness-values can not be proven.
Figure 21: Correlation lateness first two steps and last two steps
Furthermore some other additional research is done. This can be found in appendix VI.
4.4.4 Interview with planners
After the results concluded so far it seems there is a capacity shortage continually, especially for the operation or lymph node dissection. But also two interviews are processed in order to gain more insight in the planning procedure.
Last scan until diagnosis II
The waiting time between the scan and diagnosis II is influenced by the time analysts have to examine the outcome of the scan and the capacity for a specialist or nurse to discuss the results with the patient during diagnosis II. According to the planner of the radiology
department, there is no delay in examining the outcome, often the results are known the same day and therefore the delay must be assigned to the capacity for diagnosis II appointment. However, according to stakeholders the results have to be authorized which requires more time than just one day. This should be further investigated.
Diagnosis II until operation or lymph node dissection
According to the planner of the urology department, patients are planned for an operation or lymph gland dissection based upon doctor’s advice. During this interview we found out the planners do not take into account the target waiting times in this planning process. The question is, if doctors do take into account these targets, because if they don’t the problem is maybe not insufficient capacity, but causes for bad performance can be found in
communication about these target times.
One of the suggestions of the planner of the urology department for delays in the last step, is that it can be caused by the appointment with the anaesthetist, which take place before the operation or lymph node dissection (see appendix I), however according to stakeholders the
-20 -10 0 10 20 30 40 50 60 70 -20 -10 0 10 20 Late n e ss l ast scan u n til o p e ration o r ly m p h n o d e d issec tion
order in which planning is made for the appointments after diagnosis II, is first the
5.1 Introduction
This research project is focused on the reliability of patient waiting times between consecutive hospital visits within a care-path in hospital X in the north of the Netherlands.
For analysis of this care-path, data has been analysed from 2010 and 2011. The first aim was to establish to what extend real waiting times of patients comply with the criteria of standard waiting times. The second aim was to find causes for the percentage of patients not complying with target waiting times in order to find opportunities for improvement. In order to reach these goals, the framework developed by Soepenberg et al (2012) is used, which was originally developed for diagnosing the delivery reliability performance of make-to-order companies. Until now, this framework has never been used in a care-path setting. Since similarities exist between make-to-order companies and a care-path with respect to the process, an additional investigation was focused on whether this framework can be used in this setting. This chapter is focused on answering the research questions. Additionally, the usefulness of the framework will be discussed.
5.2 Research question I
To what extent do waiting times of patients in the care path meet the prescribed target waiting times?
The investigated steps of the care-path are as follows: 1: Biopsy – diagnosis I
2: Diagnosis I – first scan 3: Last scan – diagnosis II 4a: Diagnosis II – Operation
4b: Diagnosis II – Lymph node dissection
The first research question is answered by determining the distribution of lateness for every step which resulted in the percentages of complying patients. These percentages are given in bold in table 6.
Tabel 6: Results research question 1
Step Target waiting time Average waiting time Percentage of complying patients
1 10 8,73 76%
2 10 6,84 89%
3 7 8,23 47%
4a 14 24,81 15%
4b 14 28,25 0%
After answering the first research question, a discussion with the stakeholders took place. Since there was no target set for the percentage of complying patients, the question raised was what number of percentage should be sufficient. The stakeholders would be satisfied when the percentages should reach an amount of 80. According to the results, in particular the
diagnosis II until the operation or lymph node dissection. Therefore, little attention is given to the first two steps for the second research question.
5.3 Research question II
How can waiting times of patients comply with target waiting times?
For every step several options are investigated for improving the reliability performance with respect to waiting times. In this section conclusions following from this research are given and if possible, options are given for a better performance of waiting times reliability.
Biopsy until diagnosis I
For this step there is checked whether applying a FCFS rule would lead to an increasing percentage of patients complying with the target waiting time. If the FCFS rule would have been applied, there would not be a real improvement in percentages of complying patients. An option for increasing the reliability performance is making good decisions with respect to capacity. However, this is not examined because the focus of this research is on the last two steps.
Diagnosis I until first scan
For this stage, the same conclusions are made as for the first stage. A FCFS rule would not lead to higher percentage rates. Also here an option for increasing the reliability performance is making good decisions with respect to capacity. Since the performance of this step was already 89%, which is very high, no opportunities for improvements are really necessary.
Last scan until diagnosis II
* The target time between diagnosis I and diagnosis II is seventeen days, however with a maximum of twenty days. It is not obvious how to divide the three days between the two steps, within the research this maximum is not taken into account. Noteworthy is that when these three days are added to the target waiting time of this step, the percentage of meeting the target time will increase to 81%, which is a large difference of 47%.
* During the holiday seasons there is no increased lateness. So, it can be stated that the dependency between lateness and holiday periods is low for this step.
* The waiting time of this step is influenced by the time analysts have to examine the outcome of the scan as well as the capacity for a specialist or nurse to discuss the results with the patient during diagnosis II. According to the planner of the radiology department, there is no delay in examining the outcome, because often the results are known the same day and therefore the delay must be assigned to the capacity for diagnosis II appointment. However, according to the stakeholders the results have to be authorized which requires more time than just one day. To find an opportunity for increasing the reliability performance, this could be examined.
Diagnosis II until operation or lymph node dissection
* Lateness is not increased in periods of holidays, despite reducing of capacity during holidays. Possibly prostate cancer patients, do not suffer from this reduction of capacity, because they are getting priority compared to other patients.
* According to the planner of the urology department, patients are planned for an operation or lymph gland dissection based upon doctor’s advice. The planners of the urology department are not aware of the target waiting times. The question is, if doctors do take into account these targets. If they don’t, this is a possible cause for bad performance. Otherwise causes are in all probability related to insufficient capacity.
* One of the suggestions of the planner of the urology department for delays in the last step, is that it can be caused by the appointment with the anaesthetist, which take place before the operation or lymph node dissection (see appendix I), however according to stakeholders the order in which planning is made for the appointments after diagnosis II, is first the
operation/lymph node dissection and thereafter the preliminary inquiries (which includes the appointment with the anaesthetist). Therefore, in their opinion the preliminary inquiries do not affect delays. It seems this procedure is not clear to involved personnel.
Biopsy until operation or lymph node dissection
Also there is investigated if there exist a relationship between the lateness from biopsy until scans and the lateness from the last scan until operation or lymph node dissection. Only 5% of the total variation in the lateness of last scan until treatment can be explained by the linear relationship between the the first two and the last two steps. Therefore, a relationship between the lateness from biopsy until scans and the lateness from the last scan until operation or lymph node dissection can not be proven. This means, patients that get priority in the first stages do not necessarily get prioritized in the last stages.
5.4 Methodology
A great number of problems and causes for problems could be identified by using the framework of Soepenberg et al (2012). Especially because in a care-path there is a clear defined patient group, there are clear defined steps and clear defined target waiting times, analysis by using this framework was successful. However one of the drawbacks was, there was sometimes too less data available, in particular of patients who went for an operation or lymph node dissection. Therefore it was difficult to make firm conclusions. Another
