University of Michigan Health System

University of Michigan Health System Program and Operations Analysis Utilization Study of Linear Accelerators in the Radiation Oncology Department Project Report To: Kathy Lash: Director of Operations Radiation Oncology Department Ruth Neal: Allied Health Senior Supervisor Radiation Oncology Department Sheri Curnes: Senior Management Consultant Program and Operations Analysis From: Industrial and Operations Engineering (IOE) 481 Project Team, Programs and Operations Analysis - Sepehr Mowlavi: Project Team Member Zachary Shoup: Project Team Member Alex Wang: Project Team Member Date: April 16, 2007 1

Table of Contents Executive Summary 3 Introduction 4 Background 4 Key Issues 5 Goals and Objectives 5 Project Scope 5 Project Approach 5 Literature Search 5 Brainstorming Session 6 Utilization Study 6 Treatment Scheduling Study 6 Findings and Conclusions 7 Utilization Study 7 Treatment Scheduling Study 8 Recommendations 11 Expected Impact 12 Appendix A Bibliography 13 Appendix B.1 Treatment Scheduling Study Instructions 14 Appendix B.2 Treatment Scheduling Study Data Sheet 16 Appendix B.3 Brainstorming Session 17 Appendix C.1 Utilization Study Instructions 18 Appendix C.2 Partial Utilization Study Data Sheet 19 Appendix D Utilization Study Findings 20 Appendix E.1 Reasons for Late Starts 21 Appendix E.2 Reasons for Extended Treatment Times 22 Appendix F A Day in the Life of Radiation Oncology 24 2

Executive Summary The Radiation Oncology Department is an active unit that provides cancer treatments for patients Monday through Friday during the year. The Department utilizes specialized machines called linear accelerators to perform various radiation treatments ranging from Total Body Irradiation to Intensity Modulated Radiation Therapy. The operating facilities currently include four linear accelerators that perform these treatments EX1, EX2, EX3, and 600CD. Scheduling conflicts on these machines can arise due to late arrivals, emergency cases, machine limitations, and machine maintenance. As a result, the length of the working day has increased from the target of 12 hours to 15 or more. In response, the Director of Operations of the Radiation Oncology Department has requested the team s help in determining the true capacity of the linear accelerators and finding an improved way to schedule patients. The Department also seeks an estimate of the number of treatments the linear accelerators can perform each day as a benchmark for state regulators. To identify opportunities for improvement, the team completed a background study, performed a utilization study of the linear accelerators, studied the treatment scheduling process, and compared scheduled and actual treatment lengths. The team made the following conclusions: Setup time presents greatest opportunity for reduction. Machine Down time is unpredictable, but not significant. Starts were scheduled for longer than they actually lasted. Treatments are more likely to start late than run long. The team recommends the following: Implement standard work, particularly for setup and teardown processes. Maintain scheduling treatments in 5-minute blocks. Schedule new starts in shorter appointments. Establish policies and procedures for communicating with other departments. Promote flexibility in machine treatment ability and RTT treatment knowledge. 3

Introduction The Radiation Oncology Department at the University of Michigan Health System is a medical department that performs radiation therapy on cancer patients using machines called linear accelerators. Difficulties in scheduling patients due to late arrivals, emergency cases, machine limitations, and machine maintenance have increased the length of the working day from the target of 12 hours to 15 or more. The Director of Operations of the Radiation Oncology Department has requested help in determining the true capacity of the linear accelerators and finding a better way to schedule patients. The department also seeks an estimate of the number of treatments the linear accelerators can perform each day as a benchmark. This estimate can be used to compare treatment capacities with other hospitals and as a performance measure. To address these issues, the team completed a background study, performed a utilization study of the linear accelerators, studied the treatment scheduling process, and compared scheduled and actual treatment lengths. This report presents the results of our project and our recommendations to improve the Radiation Oncology Department. Background The Radiation Oncology Department at the University of Michigan Hospital is an active unit that provides cancer treatments for patients Monday through Friday during the year. Types of treatments range from conventional treatments to complex specialty treatments. On average, each linear accelerator treats about 30 patients each day. The operating facilities currently include four linear accelerators that perform radiation treatments to patients. Three of the four linear accelerators are EX machines (EX1, EX2, and EX3), and the fourth linear accelerator, a 600CD machine, is used for treatments that require a 6 MV X-ray beam. Treatments performed by the three EX machines can differ from the treatments performed by the 600CD machine, and treatments performed by one EX machine can differ from treatments performed by another EX machine. These differences result from the limited number of available supplementary equipment and/or the software needed for certain specialized treatments. Currently, first-time patients are automatically scheduled for one hour to accommodate for preliminary assessment of the patient, paperwork, questions, and familiarization of the patient with the treatment process (a Start ). Although schedules are planned in advance, several types of scheduling conflicts arise. These problems can be caused by extended treatment times of the previous patient, machine downtime where the linear accelerator is being maintained or repaired, or the unpredicted arrival of an emergency patient. Additionally, patients are usually on the same scheduled machine throughout treatment, but these issues may force the staff to move patients to another machine. Due to these circumstances, extended work hours are often required and cause daily work hours to fluctuate. The Radiation Oncology Department requested the team s help in determining the true linear accelerator capacity and finding a more efficient way to schedule patients. 4

Key Issues The following key issues drove the need for this project: Frequent late starts and treatments that last longer than their scheduled length are suspected to be caused by inefficient scheduling. The staff frequently works past normal operating hours to 8 pm and would like to shorten working hours to 7:00 am - 7:00 pm or earlier, if possible. The department would like to know where they stand regarding utilization before the addition of a fifth linear accelerator in June 2007. Inconsistent methods of recording actual treatment times, machine utilization, and downtime have made it difficult to determine the true capacity of the linear accelerators. The department lacks an estimate of capacity that considers the varying complexity of treatments to present to UMHS administrators and state regulators. Goals and Objectives The following goals and objectives were identified as steps toward providing recommendations to address the key issues faced by the Radiation Oncology Department. Establish the current capacity of the department and the utilization of the linear accelerators. Determine the frequency and primary causes of late treatment starts. Determine the frequency and primary causes of extended treatments. Determine whether treatments are being scheduled for the correct amount of time. Project Scope This project examined the scheduling and utilization of the four linear accelerators (EX1, EX2, EX3, and 600CD) in the Radiation Oncology department. Processes that occurred before the patient begins his or her treatment regiment such as simulation and dosimetry were not studied, nor was patient flow or other administrative functions of the department. The project did not analyze the treatment procedures. Instead, the study broadly defines the phases of an appointment in terms of setup, treatment, and teardown. Project Approach The team examined several aspects of the Radiation Oncology Department to identify opportunities for improvement by performing a literature search, conducting a brainstorming session, performing a utilization study of the linear accelerators, and studying the treatment scheduling process. The following section describes the methodologies used in this project. Literature Search The team reviewed medical journals regarding patient scheduling and Heijunka load leveling. A patient scheduling article provided the team with a look at scheduling heuristics in other medical 5

facilities. Another article detailed the use of a Heijunka load leveling process in a manufacturing setting. The team found that the scheduling software used by Radiation Oncology in effect fulfills this role by providing a visual picture of the day s work, easily visible to the therapists at each machine. The team also studied a final report prepared by a previous IOE 481 team to obtain preliminary information on the Radiation Oncology Department and its processes. The methods of the following studies were influenced by several sources. The brainstorming session was based on the technique of Affinity Diagramming found in The Memory Jogger II. The design of the utilization study was influenced by an existing Programs & Operations Analysis report on the use of random beepers. Please refer to Appendix A Bibliography for a list of all sources. Brainstorming Session To identify the primary causes for late starts and extended treatment times, it was important to use language familiar to the therapists on the data collection sheets. The team conducted a brainstorming session with five therapists where Affinity Diagramming was used to stimulate and organize their input. Each participant was instructed to list six reasons for late starts on Post- It notes. The team facilitated the process by organizing the 30+ responses into categories of the staff s suggestion. Once a consensus was reached, each category was given a descriptive three to five word name, which would appear on the data collection sheets. This technique was repeated to identify primary reasons why treatments extend beyond their allotted time, which also appeared on the treatment scheduling data collection sheets. Refer to Appendix B.1 for definitions, Appendix B.2 for the Treatment Scheduling data collection sheet, and Appendix B.3 for a photograph of the brainstorming session. Utilization Study The utilization study began the week of February 5 th, 2007 and ended on March 2 nd, 2007. During the three weeks of data collection, either one of three members of the Radiation Oncology staff, the chief radiation therapist, or the two control room facilitators wore a random beeper set to go off at an average rate of 2.5 times per hour. However, from 5:30 pm until close each day, a member of the team took over the beeper. Each time the beeper went off, the person carrying it placed a tally in the appropriate box on a data collection sheet to represent the state of each of the four linear accelerators. The states were patient in set-up process, beam is on patient, patient set-up teardown, room not in use, and machine down. Refer to Appendix C.1 for definition of states and Appendix C.2 for the Utilization Study data collection sheet. Treatment Scheduling Study The treatment scheduling study began February 19 th, 2007 and concluded on March 12 th, 2007. Information about each treatment, including discrepancies between scheduled and actual treatment times and reasons for late starts and extended times, were recorded by the therapists throughout the day. Data collections sheets were placed in the control room for each linear accelerator. To support the data collected, the Radiation Oncology staff provided the team with an itemized daily log of when each linear accelerator was turned on and off. 6

Findings and Conclusions Using data gathered from the studies, the team was able to estimate machine utilization and the primary reasons for extended and late treatments, which turned out to be relatively consistent for all of the machines. Utilization Study Machine utilization was assessed across all four linear accelerators. Percentages of time each machine spent in each state are presented below in Table 1. Machine State by Machine. Moreover, machine utilization was also analyzed by day of the week, though no trends were found. See Appendix D for Table 1A. Machine States for Each Machine by Day of the Week. The percentages used to establish machine utilization and estimate capacity are those presented in Table 1. See Chart 1. Machine States for a graphical representation of the data presented in Table 1. The team has defined two specific types of utilization - room utilization and machine utilization. Room Utilization includes instances when the machine was in the Beam On, Setup, and Teardown state. Machine utilization is defined as percentage of time in Beam On state only. The average number of treatments per day during data collection was 115. Table 1. Machine States by Machine Percentage of time in state Machine State 600CD EX-1 EX-2 EX-3 Beam On 43.05% 35.04% 39.57% 37.87% Machine Down 0.00% 3.23% 1.34% 0.53% Not in Use 21.12% 14.56% 13.37% 12.53% Setup 28.34% 38.01% 38.50% 39.47% Teardown 7.49% 9.16% 7.22% 9.60% Grand Total 100.00% 100.00% 100.00% 100.00% The Machine Utilization chart found below depicts the above percentages in a graphical format. 100% 90% 80% 70% 60% 50% 40% Machine Down Teardown Not in Use Setup Beam On 30% 20% 10% 0% 600CD EX-1 EX-2 EX-3 7

Chart 1. Machine Utilization Findings The preceding information has led the team to make the following observations: Overall, the Radiation Oncology Department was utilizing their rooms 83% of the time. On average, machines collectively had their beams on 39% of the time. On average, machines collectively were in setup 36% of the time. The Beam On percentage ranged from 35% to 43%. Not In Use percentage ranged from 13% to 21%. 600CD had the highest Beam On time and highest Not In Use percentage. Conclusions Given the above findings, the team has developed the following conclusions: Setup time presents greatest opportunity for reduction. Machine Down time is unpredictable, but not significant. With the addition of a 5 th accelerator, Radiation Oncology could perform 143 treatments per day with no changes to utilization percentages. Given the current machine utilization of 83% and the observed average number of treatments performed per day (115) during this study, the team has generated a table that estimates the number of treatments Radiation Oncology could potentially perform. The table presents treatments per day for 4 and 5 linear accelerators for the following cases: at current utilization (83%), at 95% utilization, and at full utilization (100%). Table 3. Estimated Number of Treatments per Day with Various Capacity Utilizations Utilization Estimated Average Number of Treatments 4 Linear Accelerators 5 Linear Accelerators Current Utilization (83%)* 115 143 95% Utilization* 131 163 Full Utilization (100%)* 138 172 *In terms of room utilization, assuming no process changes Treatment Scheduling Study The Treatment Scheduling Study was conducted to uncover the primary reasons that treatments started late and or ran longer than expected. In general, late starts were found to be more common than extended treatment times. Appendix F includes a graphical representation of a typical day created from the data gathered in this study. It shows the discrepancies between how treatments are scheduled and how they actually occur. Primary Reasons for Late Starts Late starts occurred in approximately 47% of all treatments studied in this project. Primary reasons with their respective percentages are presented below in Table 4. Reasons Late. 8

Table 4. Reasons Late Reason Late Frequency Percentage Median Delay Mean Delay Std. Dev. Not late 759 53.1% N/A N/A N/A Previous appointment late 353 24.7% 0:19 0:25 0:36 Patient late 142 9.9% 0:24 0:31 0:35 Squeeze-In 77 5.4% 0:23 0:30 0:22 Machine down 21 1.5% 0:24 0:52 1:40 Chemo 16 1.1% 3:25 4:10 2:59 Patient seeing doctor 10 0.7% 0:28 0:33 0:25 Transport 10 0.7% 0:20 0:24 0:17 Anesthesia 8 0.6% 0:22 0:33 0:30 Other 33 2.3% Total 1429 100.0% 0:20 0:33 0:56 *Refer to Appendix E.1for a complete version of the above table. Findings The following findings were established with the above information: Nearly 47% of the time treatments ran late. The top three reasons for late starts are: o Previous Appointment Late o Patient Late o Squeeze In Previous Appointment Late occurred 25% percent of the time for all appointments. Delays associated with Chemo patients were substantially longer than the delays of other treatments types. Conclusions Given the above findings and observations made during the study the team has developed the following conclusions: Once a single appointment runs late, almost all subsequent appointments will run late. The remaining reasons for late starts are significantly less than the top three reasons. Reasons for Extended Treatment Times Unlike late starts, extended treatment times are not as common. On average, treatments ran long approximately 22% of the time. Reasons long and their frequencies of occurrence are presented below in Table 5. Reasons for Long Treatments. Table 5. Reasons for Long Treatments Reason long Frequency Percentage Median Overage Mean Overage Std. Dev. Not long 1117 78.2% N/A N/A N/A Insufficient time 88 6.2% 0:05 0:07 0:08 Unspecified (<5 min) 79 5.5% 0:01 0:01 0:00 Patient setup problems 52 3.6% 0:08 0:12 0:13 Unspecified (>=5 min) 39 2.7% 0:13 0:23 0:30 Filming 23 1.6% 0:05 0:08 0:10 Machine problems 9 0.6% 0:14 0:16 0:14 Documentation problems 6 0.4% 0:05 0:06 0:05 Other 16 1.1% Overall 1429 0:05 0:09 0:15 *Refer to Appendix E.2 for a complete version of the above table. 9

Findings With the use of the above information, the following findings have been developed: Treatments run long nearly 22% of the time. The top three reasons for extended treatment times: o Insufficient Time o Unspecified (<5 min) o Patient Setup problems Insufficient Time occurred 6% percent of the time for all appointments. Conclusions Treatments are more likely to start late than run long. Actual vs. Scheduled Treatment Lengths Lastly, the data collected in the Treatment Scheduling Study was analyzed to compare scheduled versus actual length of treatments. In this portion of the study, treatments were found to be scheduled for longer than was actually necessary. Table 6. Actual vs. Scheduled Treatment Length by Treatment is presented by treatment type. Table 6. Actual vs. Scheduled Treatment Length by Treatment Treatment Frequency Mean Scheduled Length (σ) Mean Actual Difference Length (σ) Significant? (p-value) Daily Treatment 972 00:17 (6) 00:14 (15) 0:03 Yes (0.000) IMRT 289 00:25 (7) 00:22 (10) 0:03 Yes (0.000) Start 58 00:55 (13) 00:36 (20) 0:19 Yes (0.000) OBI IMRT 48 00:20 (4) 00:18 (9) 0:02 Yes (0.044) Boost 19 00:23 (20) 00:28 (23) -0:05 No (0.313) ABC IMRT 11 00:30 (4) 00:30 (5) -0:00 No (0.784) TBI 11 01:01 (23) 00:50 (31) 0:10 No (0.345) Anesthesia 5 00:36 (5) 00:25 (10) 0:10 No (0.112) OBI 5 00:20 (0) 00:19 (5) 0:00 No (0.871) Body Stereo 3 00:65 (8) 00:49 (4) 0:15 N/A E-Beam 3 00:11 (2) 00:7 (1) 0:04 N/A ABC 2 00:22 (10) 00:19 (8) 0:03 N/A Waterbath 2 00:40 (0) 00:40 (4) -0:00 N/A OBI Cone Beam 1 00:25 (N/A) 00:26 (N/A) -0:01 N/A Given the differences of Actual vs. Scheduled treatment durations, the team performed paired t- tests to check for statistical significance. 10

Table 7. Actual vs. Scheduled Treatment Length by Scheduled Length Scheduled Length Frequency Mean Actual Length (Std. Dev) Difference Signficant? (P-value) 0:05 1 0:15 (12) -0:10 N/A 0:10 151 0:10 (12) 0:00 No (0.346) 0:15 458 0:13 (7) 0:02 Yes (0.000) 0:20 489 0:18 (10) 0:02 Yes (0.000) 0:25 37 0:21 (6) 0:04 Yes (0.001) 0:30 158 0:28 (12) 0:02 No (0.212) 0:35 12 0:17 (7) 0:18 Yes (0.000) 0:40 53 0:30 (8) 0:10 Yes (0.000) 0:45 2 0:16 (19) 0:29 N/A 1:00 62 0:41 (19) 0:19 Yes (0.000) 1:15 1 0:53 (N/A) 0:22 N/A 1:30 4 0:35 (17) 0:55 Yes (0.008) 2:00 1 0:38 (N/A) 1:22 N/A Findings The following findings were developed from the above information: The durations of Daily treatments, IMRTs, Starts, and OBI IMRTs were found to be significantly different at a 95% confidence level. The differences for Daily treatments, IMRTs, and OBI IMRTs were found to be small enough to overlook given the size of appointment scheduling blocks (5 minutes). Differences were found for other treatments but statistical significance could not be established. Statistically and substantially significant differences were found between the scheduled and actual durations for 25 minute, 35 minute, 40 minute, 60 minute, and 90 minute appointments. Conclusions Starts were scheduled for longer than they actually lasted. Most appointments are scheduled for a few minutes longer than they need to be. Recommendations Standard Work One of the key findings from the capacity utilization study states that the patient spends as much time in setup as they do with the linear accelerator beam on. Much of this is a necessary part of the treatment process itself, but there is still opportunity to make it more efficient and increase the fraction of time that the machines are in use. The first step towards improvement is to standardize the process. The team recommends that Radiation Oncology implement 5S and standardized work in all of the treatment rooms and control rooms. This will help capture variation in the process that can occasionally make treatments take longer than scheduled. 11

Shorter Appointments for New Starts Currently, all appointments for new patients are scheduled for 60 minutes (with a few exceptions) to allow enough time for filming and calibrating the treatment. The team s findings clearly show that 60 minutes is more time than necessary for most cases. The team recommends that Radiation Oncology start scheduling new starts for 45 minutes, which is plenty of time for most of the new starts the team observed. Because a start involved both the initial setup and the first treatment, the team thinks that further gains could be realized by scheduling them according to how long the treatment is actually expected to last. Appointment Length Other than new starts, the team detected statistically significant differences between the scheduled appointment length and actual treatment time for daily treatments, IMRTs, and OBI IMRTs although they were only on the order of a few minutes. There appeared to be differences for other types of treatments, but there were not enough instances to observe during the three week duration of the study to make statistical inferences. The team also found differences when analyzing the appointment durations by scheduled appointment length. The differences were statistically and substantially significant for 25 minute, 35 minute, 40 minute, 60 minute, and 90 minute appointments. The scheduling software used by Radiation Oncology allows them to schedule appointments in five minute increments which they are currently doing, so the team recommends that they continue this practice. However, the team also suggests that more detailed standards and procedures be created for deciding how long a treatment is expected to last to last. These changes will hopefully lead to fewer differences between actual and expected treatment lengths. Establish Policies and Procedures for Communicating with Other Departments The team observed substantial delays associated with patients undergoing chemo therapy, anesthesiology, or being transported from another area of the hospital. The common theme among these issues is communication with other departments. The team recommends that Radiation Oncology work together with these departments to create standardized policies and procedures to facilitate better information flow regarding the scheduling and arrival of patients. Improve Flexibility of Machines and Staff The team found that the ability of the staff to transfer patients between machines on short notice was had a very positive effect on the departments ability to maximize utilization and minimize deviations from the schedule. However, the use of squeeze-ins was limited to some extend by equipment inventories specific to certain treatment rooms and a tendency for squeeze-ins to occur within a single control room, such as transferring a patient between the EX-1 and EX-2, rather than the department as a whole. The team recommends that the department makes each treatment room as flexible as possible. This may require the purchase of some additional equipment, in which case a cost-benefit analysis may be necessary. Expected Impact As a result of this project, the team provided recommendations that will improve the Radiation Oncology Department. The team expects recommendations to lead to: 12

Shorter setup times for treatments By implementing standardized work practices the RTT and other Radiation Oncology staff will become more efficient in their patient setup processes. This efficiency will then lead to shortened setup times and ultimately improve the efficiency of treatments and minimize the patient time in setup. Reduced frequency of late appointments Promoting communication with other departments in the University of Michigan Hospital will reduce the frequency of in-house patients arriving late for appointments. Shorter waiting time for patients Lastly, the implementation of all the above changes an overall improve of processes efficiency will result and allow patient appointments to be met per schedule. 13

Appendix A - Bibliography I. Vermeulen, S. Bohte, K. Somefun, and H. La Poutre. Improving patient activity schedules by multi-agent Pareto appointment exchanging. CEC/EEE 2006 Joint Conferences, 2006. D. T. Jones. Heijunka: Leveling Production. Manufacturing Engineering. Dearborn: Aug 2006. Vol.137, Iss. 2; pg. 29, 5 pgs. K. Cresci and J.W. Kang. Analysis of Board Runner Workload for Interventional Radiology. Program and Operations Analysis, Fall 2006. M. Brassard and D. Ritter. The Memory Jogger II A Pocket Guide for Continuous Improvement & Effective Planning, 1 st edition. GOAL/OPC, 1994. 14

Appendix B.1 Treatment Scheduling Study Instructions Instructions for Treatment Scheduling Sheet IOE 481 Winter 2007 1. One Treatment Scheduling Sheet will be placed at each machine each day and must stay with its respective machine throughout the day. Relevant information for all treatments on each machine should be recorded using the steps below. 2. Circle the appropriate machine at the top of the page and record the date. For each treatment: 3. Under Treatment Type, pick the type of treatment and check the appropriate box. If the treatment is not listed, check the blank entry and write in the name of the treatment. 4. Under Appt. Time, write the time that the appointment was scheduled to begin. 5. Under Scheduled Machine, check the box next to the machine that the treatment was scheduled for. 6. Under Scheduled Length, check the box next to the length of time the appointment was scheduled for. 7. Under Time patient was greeted, record the time when the patient is greeted by RTT staff. 8. Under Time patient left, record the time when the patient leaves. 9. If the appointment began after its scheduled time ( Appt. Time column), circle Y under Late Start? Circle N otherwise. 10. If Y was circled in Step 9, check the box next to the reason for the late start. If the reason is not listed, check the last box and write a brief description. 11. If the appointment lasted longer than the scheduled duration ( Scheduled Length column), circle Y under Ran Over? Circle N otherwise. 12. If Y was circled in Step 11, check the box next to the reason for running over. If the reason is not listed, check the last box and write a brief description. Definitions for Reasons for Late Start Previous appt. ran late The preceding appointment ends after the current appointment is scheduled to begin. Squeeze-in(s) A patient scheduled on another machine is squeezed-in the schedule for the current machine due to machine downtime or scheduling conflicts. Chemo patient The patient arrives from a prolonged chemotherapy treatment. Patient was late The patient arrives at a time that prevents the treatment from starting at its scheduled starting time. Plan not ready Plans for treatment are still being developed and/or changed by the doctor. Machine down Machine cannot be utilized and is in need of maintenance. Definitions for Reasons for Running Over Waiting for physicians Additional medical staff is needed to start or check patient. Patient setup problems The patient experiences discomfort and must be readjusted or conditions change which necessitates additional filming. Special cases Patients needing specialized treatments such as anesthesiology Documentation problems Problems associated with documentation (i.e. chart missing, films were not checked by doctor). 15

Start problems Problems associated with Starts (i.e no consent, treatments with too many angles to film). Insufficient time Given no problems or complications, the scheduled amount of time was still not enough to complete the treatment on time. 16

FOR Appendix B.2 Treatment Scheduling Study Data Sheet 17

Appendix B.3 Brainstorming Session This is a photo of the responses the team received from the therapists who participated in the brainstorming session. This information was used to create the treatment scheduling data collection sheet. 18

Appendix C.1 Utilization Study Instructions Instructions for Utilization Study IOE 481 Winter 2007 1. One utilization study data sheet should be completed each day. Only one sheet in needed for all four linear accelerators. 2. Every morning, the beeper should be turned on at 7:00pm and left on until the last scheduled patient finishes their treatment. 3. At the end of the day, whoever is collecting data with the utilization study sheet should place it in Gwen Neal s office in the folder labeled IOE. Each time the beeper goes off: 4. Walk down the hallways where all four linear accelerators are housed and asked the RTT staff team responsible for each linear accelerator what state from the below states the room/machine is in at the exact moment in time. 5. When the state is given, place a tick mark in the row that shares the same name as the state collected for each machine. Definitions of States Beam On When the linear accelerator beam is on and the light above the linear accelerator room labeled BEAM ON is lit. Machine Down When a machine cannot be used do to malfunction or scheduled maintenance. Not In Use When the room is empty and no one is performing any work towards setting up for the next appointment or tearing down from the previous appointment. Setup When the room is in use and setup processes are taking place with the beam off. Teardown When the treatment is finished and processes are in place to clean and teardown any hardware associated with the previous appointment. 19

Appendix C.2 Partial Utilization Study Data Sheet Beeper Collection Sheet Date: Names: Beeper on: Beeper off: 600CD Patient in setup process Beam is on patient Patient setup teardown Room not in use Morning (7:00 am to 5:30 pm) Evening (5:30 pm to close) Machine down EX-1 Patient in setup process Beam is on patient Patient setup teardown Room not in use Machine down 20

Appendix D Utilization Study Findings Table 1A. Machine States by Day of Week and Time of Day Sum of # State Machine Day of Week Beam On Machine Down Not in Use Setup Teardown Grand Total 600CD Monday 52.94% 0.00% 13.24% 25.00% 8.82% 100.00% Tuesday 46.74% 0.00% 22.83% 23.91% 6.52% 100.00% Wednesday 36.00% 0.00% 22.67% 34.67% 6.67% 100.00% Thursday 38.67% 0.00% 25.33% 28.00% 8.00% 100.00% Friday 40.63% 0.00% 20.31% 31.25% 7.81% 100.00% 600CD Total 43.05% 0.00% 21.12% 28.34% 7.49% 100.00% EX-1 Monday 38.81% 16.42% 10.45% 25.37% 8.96% 100.00% Tuesday 32.98% 1.06% 12.77% 41.49% 11.70% 100.00% Wednesday 29.33% 0.00% 26.67% 40.00% 4.00% 100.00% Thursday 34.25% 0.00% 9.59% 46.58% 9.59% 100.00% Friday 41.94% 0.00% 12.90% 33.87% 11.29% 100.00% EX-1 Total 35.04% 3.23% 14.56% 38.01% 9.16% 100.00% EX-2 Monday 40.30% 0.00% 8.96% 43.28% 7.46% 100.00% Tuesday 38.95% 3.16% 13.68% 35.79% 8.42% 100.00% Wednesday 49.33% 0.00% 13.33% 33.33% 4.00% 100.00% Thursday 34.25% 1.37% 16.44% 41.10% 6.85% 100.00% Friday 34.38% 1.56% 14.06% 40.63% 9.38% 100.00% EX-2 Total 39.57% 1.34% 13.37% 38.50% 7.22% 100.00% EX-3 Monday 36.36% 0.00% 10.61% 43.94% 9.09% 100.00% Tuesday 42.11% 0.00% 9.47% 34.74% 13.68% 100.00% Wednesday 41.33% 0.00% 14.67% 38.67% 5.33% 100.00% Thursday 30.14% 1.37% 17.81% 42.47% 8.22% 100.00% Friday 37.88% 1.52% 10.61% 39.39% 10.61% 100.00% EX-3 Total 37.87% 0.53% 12.53% 39.47% 9.60% 100.00% Grand Total 38.89% 1.27% 15.39% 36.08% 8.37% 100.00% Sum of # State Beam Machine Machine Period On Down Not in Use Setup Teardown Grand Total 600CD Day 41.94% 0.00% 21.29% 29.68% 7.10% 100.00% Evening 48.44% 0.00% 20.31% 21.88% 9.38% 100.00% 600CD Total 43.05% 0.00% 21.12% 28.34% 7.49% 100.00% EX-1 Day 37.34% 0.00% 15.26% 37.66% 9.74% 100.00% Evening 23.81% 19.05% 11.11% 39.68% 6.35% 100.00% EX-1 Total 35.04% 3.23% 14.56% 38.01% 9.16% 100.00% EX-2 Day 40.32% 1.61% 11.29% 39.03% 7.74% 100.00% Evening 35.94% 0.00% 23.44% 35.94% 4.69% 100.00% EX-2 Total 39.57% 1.34% 13.37% 38.50% 7.22% 100.00% EX-3 Day 40.51% 0.64% 12.54% 36.01% 10.29% 100.00% Evening 25.00% 0.00% 12.50% 56.25% 6.25% 100.00% EX-3 Total 37.87% 0.53% 12.53% 39.47% 9.60% 100.00% Grand Total 38.89% 1.27% 15.39% 36.08% 8.37% 100.00% 21

Appendix E.1 Reasons for Late Starts Reason Late Frequency Percentage Median Delay Mean Delay Std. Dev Not late 759 53.1% N/A N/A N/A Previous appointment late 353 24.7% 0:19 0:25 0:36 Patient late 142 9.9% 0:24 0:31 0:35 Squeeze-In 77 5.4% 0:23 0:30 0:22 Machine down 21 1.5% 0:24 0:52 1:40 Chemo 16 1.1% 3:25 4:10 2:59 Patient seeing doctor 10 0.7% 0:28 0:33 0:25 Transport 10 0.7% 0:20 0:24 0:17 Anesthesia 8 0.6% 0:22 0:33 0:30 Plan not ready 6 0.4% 0:22 0:21 0:15 Double-booked 6 0.4% 0:34 0:29 0:15 Couldn't find chart 4 0.3% 0:10 0:11 0:05 Fire drill 4 0.3% 0:19 0:28 0:20 Other 3 0.2% 0:40 0:34 0:25 Warmup 3 0.2% 0:10 0:10 0:01 Waiting for physician 2 0.1% 0:19 0:19 0:12 hydration 1 0.1% 0:14 0:14 N/A Not aware of check in 1 0.1% 0:54 0:54 N/A Patient at reassessment 1 0.1% 1:30 1:30 N/A Patient coming from MRI appointment 1 0.1% 0:39 0:39 N/A PI 1 0.1% 0:45 0:45 N/A Total 1429 100.0% 0:20 0:33 0:56 22

Appendix E.2 Reasons for Extended Treatment Times Reason long Frequency Percentage Median Overage Mean Overage Std. Dev. Not long 1117 78.2% N/A N/A N/A Insufficient time 88 6.2% 0:05 0:07 0:08 Unspecified (<5 min) 79 5.5% 0:01 0:01 0:00 Patient setup problems 52 3.6% 0:08 0:12 0:13 Unspecified (>=5 min) 39 2.7% 0:13 0:23 0:30 Filming 23 1.6% 0:05 0:08 0:10 Machine problems 9 0.6% 0:14 0:16 0:14 Documentation problems 6 0.4% 0:05 0:06 0:05 Other 4 0.3% 0:23 0:28 0:18 Waiting for physician 4 0.3% 0:03 0:07 0:09 Start problems 2 0.1% 0:35 0:35 0:31 Doctor came on set 1 0.1% 0:01 0:01 N/A Had to take all SSD's 1 0.1% 0:03 0:03 N/A New plan 1 0.1% 0:22 0:22 N/A PI 1 0.1% 0:09 0:09 N/A Special cases 1 0.1% 0:16 0:16 N/A Verifiying ISO 1 0.1% 0:08 0:08 N/A Total 1429 100.0% 0:05 0:09 0:15 23

Appendix F A Day in the Life of Radiation Oncology 24