Process optimization at the outpatient
clinic of the Clinical Immunology and
Rheumatology department
A Lean Six Sigma project
-Marieke Backer, student no 10733647, mariekebacker@gmail.com, Master thesis for the UvA MBA special track in Health Care Management.
Supervisors:
Dr. Marit Schoonhoven, Associate Professor at University of Amsterdam; Senior Consultant IBIS UvA Management; m.schoonhoven@uva.nl
Dr. Marieke van Onna, Rheumatologist; Clinical Immunology and Rheumatology department at AMC; m.vanonna@amc.uva.nl
-2-
Executive Summary
The Clinical Immunology and Rheumatology (KIR) department of the Academic Medical Centre (AMC) deals with immune mediated diseases and musculoskeletal complaints. The department consists of 10 staff members (rheumatologists), 8 residents (rheumatologists in training) and 4 rheumatology nurses also called (known as) verpleegkundig reumaconsulten (VRC) who together perform 11.5 full-time equivalent (FTE) on tasks in the outpatient clinic. The processes at the outpatient clinic have changed significantly in recent years, mainly due to the implementation of a new Electronic Health Record (EHR): EPIC. As a consequence, the employees need more time to fulfill all the tasks related to the outpatient clinic resulting in less time for other tasks like research, management and education. The patient planning at the outpatient clinic is not optimally adjusted to these new processes either, which is resulting in longer waiting times for the patient.
The objective of this project is to optimize the time that all employees of the KIR outpatient clinic spend on patient care at the outpatient clinic.
This project makes use of the Define, Measure, Analyse, Improve, and Control (DMAIC) approach (De Mast 2012) to decrease consultation time, time spent on other outpatient clinic tasks such as meetings, administration and consultation with colleagues, and waiting time for patients. For almost 30% of the patients the waiting time is 10 minutes or longer. The ultimate goal is to reduce the time spent on outpatient clinic related tasks and
consultation with 10% (1 FTE) and to achieve waiting time no longer than 10 minutes. The study reveals that during outpatient care tasks, the employees spend only 31.5% of their time on outpatient clinic consultations and 68.5% on other tasks. The most potential influential factor on the time spent on other tasks is time spent on completing
administrative tasks. From the 7.9 FTE spend on other tasks, 31.1% (2.45 FTE) is spend on administrative tasks. A reduction of 50% of the administration time will therefore save 1.2 FTE. During the consultations the employees also spend 50% of their time on administrative tasks. A 50% reduction of this time would save 0.9 FTE during consultations and another 0.7 FTE during telephone consultations. The most important influence factor on administration is the new EHR, which needs to be adapted to the work flow of the outpatient clinic. In total, the potential improvement of this project is 2.8 FTE. This result shows that optimizing the processes by means of a Lean Six Sigma project could lead to more efficient use of employee time (i.e. less time spent on waste tasks,more time for interaction with patients and other academic tasks), and less waiting time for patients.
-3- Table of content
Executive Summary ... 2
I. Introduction ... 5
A. Academic Medical Centre ... 5
A. Department of Clinical Immunology and Rheumatology (KIR) ... 5
B. Staff ... 6
C. The Problem ... 7
D. Workload ... 7
E. Improvement in hospitals ... 8
F. Goal of the project ... 9
II. Method ...10
A. Six Sigma ...10
B. Lean ...10
C. Lean Six Sigma ...11
D. DMAIC ...12
III. Results ...14
A. Define ...14
DMAIC 0 ...14
B. Measure ...17
DMAIC 1: Define the CTQs ...17
DMAIC 2: Validate the measurement procedures ...19
C. Analyze ...19
DMAIC 3: Diagnose the current process ...19
DMAIC 4: Identify potential influence factors ...26
A. Improve ...27
DMAIC 5: Establish the effect of influence factors ...27
DMAIC 6: Design improvement actions ...35
B. Control ...42
DMAIC 7: Improve process control ...42
-4-
IV. Critical reflection, Conclusion and recommendations (non-technical) ...46
A. Critical reflection ...46 Disruptive solutions ...46 B. Recommendations ...48 C. Conclusion ...49 V. References ...50 VI. Appendices ...54
Appendix I travel sheet form ...54
Appendix II DILO form ...55
Appendix III list with all potential influence factors that results out of the Gemba walk ...59
Appendix IV ANOVA test ...62
Appendix V phone call routing schedule ...64
-5-
I. Introduction
A. Academic Medical Centre
The Academic Medical Centre (AMC) in Amsterdam is one of the eight academic centers in The Netherlands. Besides patient care, research and training are the other two core activities of the AMC. Within the AMC the medical departments are organized into divisions. Other functions are organized through services, support groups and directorates. Division A is the department of internal medicine in the AMC (Figure 1). The department of Clinical
Immunology and Rheumatology (KIR) is a sub-department of this department.
A. Department of Clinical Immunology and Rheumatology (KIR)
The KIR department deals with immune mediated diseases and musculoskeletal complaints. The European Association of Rheumatology (EULAR) and the International Federation of Immunology Societies (FOCIS) have appointed the KIR department as a center of excellence. (Website AMC).
-6- Figure 1: Organogram AMC
B. Staff
The KIR department consists of 10 staff members (rheumatologists; in total 2.8 FTE), 8 residents (rheumatologists in training; in total 6.4 FTE but varies over time) and 4
rheumatology nurses (in total 2.3fte) also called verpleegkundig reumaconsulten (VRCs). VRCs are specialized nurses in rheumatologic care. In total there are 32 consulting sessions (24 physicians and 8 VRCs sessions). Each consulting session lasts 4 hours. Within their daily duties, the staff members combine patient care at the outpatient clinic with various other activities such as care for clinical patients admitted to the hospital, management, research, education and supervision of the residents.
-7-
Only 7 of 10 staff members work in the outpatient clinic and the 2.8 FTE in total is spread over these staff members in a highly variable degree. The eight residents perform about 80% of their time on patient care related activities (6.4 FTE in total at the moment). The rest of their time they use for formal educational activities. The 4 VRCs work 2.3 FTE in total at the KIR outpatient clinic and are not employed by the KIR department, but managed by Poli P. This separate division manages all processes and logistics at the outpatient clinic. All employees with supporting tasks are deployed by Poli P, including the employees at the front office (Q2) and the back office (Q3) of the outpatient clinic who facilitate the KIR outpatient clinic are employed by Poli P as well.
C. The Problem
The processes at the outpatient clinic have changed significantly in recent years. The
reorganization of Poli P and the introduction of the new Electronic Health Record (EHR) EPIC had major consequences for the processes. In addition, the AMC received the Joint
Commission International (JCI) accreditation. The JCI accreditation program aims to improve patients´ safety and quality of health care. As a consequence, many employees have the experience that they spend less and less time on direct patient care. In addition, the planning of outpatient clinic visits is not adjusted to the new processes, resulting in longer waiting times for the patient.
D. Workload
The complex administration around patient care is performed partly by the physicians. Compared to previous practice, doctors have to enter more detailed data in the EHR during their contact with the patients. Due to all of these additional administrative tasks they experience a much heavier workload and have to spend more time on this than before. In general, according to 92% of physicians the administrative work for professionals has increased since 2000 (Plexus 2010). The administrative costs of hospitals in the Netherlands accounts for 19.8 percent of all hospital costs, which is less than in the US (25.3 percent), but much more compared to France and Germany (both 11.6 percent). The high administration costs of the Dutch hospitals are driven by the complexity of the reimbursement system and
-8-
the accompanying heavy regulatory burden. Moreover, when total hospital costs in different countries are compared it is clear that nations with the highest hospital administration costs have also the highest total hospital costs (Himmelstein 2014). Of note, in the US a quarter of the administration costs is superfluous and could be saved (Cutler 2012).
A study of Menzis shows that, on average, 53% of the time in health care is spend on care tasks and 47% on indirect tasks (Menzis 2010). There are no exact numbers known for specialists in hospitals, but general practitioners spend on average 11 to 13.5 hour each week on administration (Nivel). According to Weening et al., an employed medical doctor in general spends 13% of their time on administrative tasks (Weening 2002).
E. Improvement in hospitals
In general, processes in hospitals are evaluated in time and have to be redesigned, which is also the case in the AMC. The current processes result in unnecessary duplications of
services, long waiting times and delays. The processes also do not meet the patients’ needs. Furthermore, the inefficient processes may cause medical errors, time consuming rework and therefore result in substantial financial losses (Feigenbaum 1951).
In operational management terms, a hospital consists of all different kinds of medical and non-medical processes for diagnostics and treatment of the patient (de Mast 2012). The concepts of process speed and lead time have direct influences on costs in most industries, where the customer is only interested in the end product. But in healthcare patients experience the whole process and are not only interested in the end product. So lead time, waiting times and waste have a direct impact on the quality perception of patients (van de Heuvel et al. 2006). The patient is part of the process and an imperfect process results in defects and rework which affect the patient´s safety directly.
According to Peter Drucker, employees are the most important resource in hospitals and hospital management should invest in increasing the productivity and knowledge of employees (Drucker 1974). The nurses and doctors know more about the specialized fields and the needs of the patient than the management does. Lean is a method that aims for a balance between workload and capacity of all resources, including the employees. The idea
-9-
is that employees work more efficiently when waste, inefficiencies in the process and disorders are eliminated (Does 2015).
Lean Six Sigma is successful in healthcare. It is used to reduce costs (de Mast 2012) and to improve the quality of care and organization ( Kennedy 2010). By implementing Lean, the hospital becomes a learning organization. The knowledge of the employees is shared and improvements are implemented together with the employees. This motivates the employees and keeps them engaged. Lean Six Sigma has been implemented in hospitals in the
Netherlands (de Koning 2006). In 2001, the Red Cross Hospital was one of the first Dutch hospitals to implement Six Sigma which resulted in €1.2 million in annual savings (van den Heuvel, 2005b). The Commonwealth Health Corporation is one of the first healthcare organizations that implemented Six Sigma. The implementation in 1998 resulted in savings of $2.5 million (Lazarus and Stamps 2002b).
F. Goal of the project
The goal of this project is to optimize the time that employees of the KIR outpatient clinic spend on outpatient clinic related tasks. When the time on outpatient clinic tasks is optimally spent, employees will have more time for other duties and for the patients. In addition, the planning in the outpatient clinic should be adjusted to the new processes , which will result in shorter waiting times for the patient.
In the next chapters the used method, Lean Six Sigma, is explained. It describes the origin from Lean and Six Sigma and the benefits of combining these two methods in the Lean Six Sigma approach. In chapter three the results of the project are described according the 8 DMAIC steps. It describes the current situation, the most important influence factors and corresponding improvement actions. The last chapter concludes the project with a critical reflection and conclusion of the project and recommendations for the future.
-10-
II. Method
A. Six Sigma
By definition, Six Sigma is an organized and systemic method for strategic process
improvement and new product and service development relying on statistical (methods) and scientific methods in order to make dramatic reductions in customer defined defect rates (Linderman 2003).
In 1987, Motorola was the first company that used Six Sigma to organize its process
improvements. But the real boost for Six Sigma came with General Electric started using it in 1995, after which this method got adopted by many other companies (Does 2014).
Six Sigma is a scientific method which helps solving problems in industry and business. The methodology focuses on measuring defects, reducing variation, improving quality of processes, products and services (Does 2015). It is characterized by prioritizing cost
reduction, customer-driven approach and decision making, based on analysis of quantitative data. The focus on data has proven to add significant value to the more subjective and intuitive thinking in improvement. The cycle of DMAIC (Define-Measure-Analyze-Improve-Control) can resolve quality problems and offers an organizational structure to implement and retain improving changes (Harry 1997). The weakness of Six Sigma is its complexity and lack of standard solutions (de Koning 2006). The term “sigma” defines the standard deviation of a random variable. A Six Sigma process indicates that only 3.4 defects per million
opportunities are likely to occur (van den Heuvel 2005a).
B. Lean
‘A Lean process is a process without waste, that delivers efficiently and smoothly what the customer needs.’ (Does 2015).
Lean was introduced into the Western world in 1990 with the publication ‘The machine that changed the world’ (Womack, Jones and Roos 1990). The basis of Lean in the Toyota
Production system (TPS) is closely related to the Just-In-Time (JIT) principles (Ohno 1988, Shingo 1989, Shah 2007).
-11-
Reducing waste, managing variability and synchronizing process flows are the main focus points in Lean. The framework of Lean helps to analyze the processes within the organization with an integrated system of principles, tools and techniques (Standard and Davis, 1999). The framework is based on five organization principles: value, value stream, flow, pull and elimination of waste (Womack 1991). There are three types of waste: Muda, Muri and Mura. Muda stands for non-value adding tasks. By getting rid of the non-value adding tasks
employees can finish more tasks without working harder. Muri is the overuse of staff, systems and equipment and Mura represents the unplanned variations in a process. Lean works with a set of standard solutions like line balancing, 5S- method, one piece flow, pull systems, one piece flow and visual management (Standard and Davis 1999; George 2003).
Kaizen is one of the main tools of Lean. With Kaizen, which is Japanese for continuous improvement, employees take the initiative to improve their process. This makes sense given the fact that the employees know the process by heart and are able to recognize the smallest details crucial for process improvement. The bottom up approach makes sure that employees are committed and changes are more easily accepted. Kaizen works particularly well for small changes. For big changes, on the other hand, a top down approach is more suitable. There is a vast number of stakeholders involved in the changing process which makes it hard to find a solution that makes everyone happy.
C. Lean Six Sigma
After 2002, a combination of the Lean production and Six Sigma resulted in the Lean Six Sigma approach (De Mast2012).
With complementary benefits, Six Sigma offers the management structures and analyzing tools, while Lean provides the tools for improvement by eliminating waste, organizing and simplifying the work processes. The operational and cultural changes lead to a total chain improvement (Pepper 2010). Lean Six Sigma provides a detailed method using standard analysis tools and techniques to reduce variation in processes, making use of performance metrics and the use of statistical tools. (de Mast 2012)
-12-
Lean Six Sigma projects are conducted by Black or Green Belts, people who are trained in the Lean Six Sigma method, statistical tools and techniques for problem solving (Pyzdek 2001). The Lean Six Sigma framework should be aligned with the organization strategy (Porter 1980) and supported by the top management (Blackmore 2013). The aligned framework results in continuous process improvement activities.
D. DMAIC
The cycle of DMAIC is used for project approach in the Lean Six Sigma method.
Define
In the Define phase the problem that the project aims to solve is described and a benefit analysis is given (De Koning 2006). In the project charter, the project leader, process owners, scope, timeline and additional members of the project are determined. A SIPOC (Supplier, Input, Process, Output, Customer) is made to clarify the problem. In a stakeholder analysis, the stakes and influences of the people that are involved in the process are analyzed.
Measure
In the Measurement stage the problem is translated into a measurable form and the current process is measured. To quantify the process performance, so called critical to quality (CTQ), indicators are determined. In a CTQ flow down the performance indicators are linked to strategic focus points of the organization. (Niemeijer 2011). Methods that can be used in the Measure phase are a travel sheet or a Day-In-the-Life-Of (DILO) measurement
Analyze
In the Analyze phase the data from the measurements are used to diagnose the current process performance and factors that potentially influence de CTQ behavior are identified (Mast 2012). Methods to discover the most important influence factors are the Gemba walk and Failure Mode and Effect Analysis (FMEA) method. Gemba means workplace and refers to the place where the real value is created in an organization. During the Gemba walks, the current process is structured in comparison to the perfect process. With the FMEA method the causes and effects of each disturbance are determined and prioritized.
-13-
Improve
In the Improve stage, the most important influence factors, i.e. influence factors that have the largest impact on the CTQ, are selected. The discovery of causal influence factors are also known as leverage variables or the X’s (Hahn 1999). For each influence factor an improvement action is developed. The scientific ‘data driven’ approach is helpful in dealing with resistance during the implementation of the results.
Control
With the improvements in place, the process will be actively monitored so that the results of the improvement will sustain.
-14-
III. Results
A. Define
DMAIC 0
The deliverables of this stage are: SIPOC analysis, benefit analysis, CTQs (Critical to Quality) definitions, project organization, stakeholder analysis and a process flowchart.
The aim of this project is to optimize the time spent on outpatient clinic tasks by the employees of the KIR outpatient clinic and reduce the waiting time for the patient. The scope of the project will cover all tasks that are done in the outpatient clinic or directly related to the outpatient clinic tasks of all VRCs and physicians of the KIR department. Other employees in the outpatient clinic who are employed by the department poli P, such as the assistants on the phone (back office Q3) or patient desk (front office Q2), as well as the other tasks, such as education and research of the physicians and VRCs, are out of the scope of the project.
SIPOC stands for Suppliers, Inputs, Process, Outputs and Customer and is used to summarize the process that will be analyzed (Figure 2). The suppliers for the process are the medical specialists and general practitioners who refer patients to the KIR outpatient clinic. The input for the process are the patients. The process involves the treatment and care of these
patients in the outpatient clinic. The process starts when the patient enters the outpatient clinic. The physician calls the patient into his/her office to perform the anamnesis, a physical exam and orders new tests when needed. The physician sets a diagnosis and starts, if
necessary, a new treatment. In patients suffering from rheumatoid arthritis a new treatment could involve a change in medications or a direct injection in the affected joint. The policy is discussed with the patient and a new appointment is made. The majority of the patients at the outpatient clinic have chronic diseases and therefore return to the outpatient clinic on regular basis.
-15- Figure 2: SIPOC of the outpatient clinic process
The benefit of this project is the reduction of time spend on non-value adding tasks, time that can be spend on direct value adding care of patients, or other tasks like education and research. By using the time in the outpatient clinic more effectively, both the employees and patients will feel more satisfied. The employees of the KIR department spend in total 11.5 FTE on outpatient clinic tasks (rheumatologists 2.8 FTE; residents 6.4 FTE; VRCs 2.3 FTE). According to the rheumatologists, in daily practice, the time spend on outpatient clinic related tasks seems to be higher. VRCs spend 38% (32 hour) of the 2.3 FTE on consultations. The physicians spend 29% (96 hour) of the 9.2 FTE on consultations. The aim is to reduce the time spent on outpatient clinic related tasks and consultations with 10%, resulting in a time reduction of 1 FTE in total.
In the outpatient clinic, the tasks of the VRCs and physicians are more or less similar: preparing the consultations, performing anamnesis, answering questions regarding their disease and treatment, checking the medication in the EHR, performing a physical exam or a joint score, explaining the policy and the plans for treatment to the patients, giving the patients advise about their health and complete the administration in the EHR. The VRCs and physicians patient populations differs. The patients that consult the VRCs are a well defined group with clear diagnosis and demands for care. The patients that consult the physicians are more complex and patients that visit the outpatient clinic for the first time consult the physician for a diagnosis and treatment plan.
Another goal is to reduce the waiting time of patients before their appointment. The waiting time is an important quality aspect for patients, the customers of the outpatient clinic.
SIPOC
Suppliers Inputs Process Outputs Customers
Own rheumatologist (returnal appointment) patient patients treatment patiënt under treatment patients
Referring physicians and care in stabilized disease
1. medial specialis outpatient clinic
2. general practinioners
Step 1 admission
Step 2 Anamnesis
Step 3 Physical exam and treatment
Step 4 Policy and advise
Step 5 planning a new appointment
-16-
All suppliers and customers are identified as project stakeholders and taken into account in the stakeholder analysis. The main stakeholder is the medical team of the KIR department consisting of rheumatologists, residents and VRCs. Other important stakeholders are the patients and the department poli P.
The team for this project consists of staff members, residents and VRCs of the KIR department. One of the staff members responsible for the planning of the consulting sessions acts as Champion (Figure 3).
Figure 3: Project organization
The micro process description differs between patients, employees and consulting sessions, although, in general, the process looks like the process depicted in figure 4.
Supplier Champion User MBB
Head of the KIR department Staf physician Head of the KIR department Marit Schoonhoven (the person supplying
resources such as time and budget)
(the person who owns the problem)
(the person who reaps the benefits from the project)
(expert in Six Sigma methodology)
Black belt Green belt Green belt Green belt
Marieke Backer Resident VRc Staf member
Investment in time (hrs./week) Investment in time (hrs./week) Investment in time (hrs./week) Investment in time (hrs./week) 8 hours 1 week 1 week 1 week
Team members
Staff VRCs residents
-17- Figure 4: Micro process
B. Measure
DMAIC 1: Define the CTQs
In DMAIC 1, the CTQ are defined and operationalized.
The first CTQ measures how the time of the employees is distributed over the outpatient clinic tasks. The tasks include consultation sessions and telephone consultations,
administration and meetings during and outside the consultation session. Research and training activities are tasks that are not related to the outpatient clinic and are out of scope of the project.
The second CTQ is the time spent on each consultation. Per consultation, the time spent on value adding and non-value adding tasks have been measured during consultation session for the rheumatologist, residents and VRCs.
The third CTQ is the waiting time for scheduled patients at the outpatient clinic. The waiting time for unscheduled patients or the time patients spend waiting on the phone is not taken into account. The defined CTQs are depicted in an adapted CTQ flow down (Figure 5) of the general CTQ flow down of Nijmeier et al. (2011) which shows an increase in revenue by increasing capacity.
-18- Figure 5: CTQ flowdown
Measurement plan
The CTQs are operationalized in the measurement plan (Figure 6).
Figure 6: Measurement Plan
In order to obtain information about the time employees spend on different outpatient clinic tasks, the activities of eight employees were measured for one week by a DILO (see
appendix II for DILO form). This has resulted in 24 registered days either by a physician or a VRC.
During the consultation sessions, the activities of the physicians and VRCs were shadowed by the Black Belt in order to obtain data about the consultation time. For this purpose a travel sheet has been used (see appendix I for travel sheet form). In total, twelve consultation sessions of each four hours were analyzed and the adjacent hours were measured by
self-Measurement plan
CTQ Unit (per what? per
job? per request? per hour? per day?)
Measurement procedure Requirement Sample size
Who will collect the data?
Outpatient clinic related tasks Minutes per day per employee
DILO As low as possible 24 days Employee Waiting time patiënt Minutes per patiënt travelsheet 0 minutes 63 Black Belt Consult time Minutes per consult travelsheet As low as possible 55 Black Belt
-19-
registration. During this measurement, the data of 64 consultations (out of which 55 control patient consultations were carried out with a physician) and 47 telephonic consultations were obtained.
From the travel sheets of the consulting sessions, the waiting time for each patient was collected by comparing the time of entering into the consulting office to the appointment time.
DMAIC 2: Validate the measurement procedures
The CTQ time spent on outpatient clinic related tasks was measured by the employees. All employees were instructed by the Black Belt in order to ensure a consistent measurement procedure. After collecting the data, the data have been checked for outliers and they were analyzed in detail. All data was checked to be complete and to make sense. During the both measurement periods no severe problems did occur.
The measurements of the CTQ consultation time and waiting time were all done by the Black Belt to avoid interpersonal measurement errors.
C. Analyze
DMAIC 3: Diagnose the current process
CTQ 1 outpatient clinic related tasks
In total, 8 employees measured their activities with the DILO method. This resulted in data of 24 days and in total of 181 hours. This is more than 7,5 hours a day.
The employees collected data on the tasks they were doing and how much time they spent on the task. The tasks are tagged as ten activities:
1. consultation sessions (consultation time)
2. preparing and completing administration of the consultation sessions 3. answering email and InBasket messages
-20-
5. consultation by phone (patients and colleagues) 6. waiting time
7. interruptions (pagers, phone calls, a colleague that interrupts) 8. meetings
9. other tasks related to the outpatient clinic 10. other tasks not related to the outpatient clinic
The tasks were divided in four main categories:
1. consultation - the hours spend on consultation sessions at the outpatient clinic or TC consultations
2. consultation other - all other tasks related to the consultation sessions like administration and short consultations with colleagues
3. other - other tasks like meetings
4. waste - like waiting time and interruptions
Figure 7: Pie chart of the four main categories of tasks performed during the DILO measurements Other Consultation other Waste Consultation Category 31,8% 4,2% 40,9% 23,0%
-21-
This shows that 31.8% of the employees’ time is used for consultation sessions, 40.9% for all other tasks related to the consultation sessions and 23.0% for other tasks (Figure 7).
This means that 31.8% of 11.5 FTE is used for consultation time and 68.5% for other tasks. This equals 3.7 FTE spent on consultation time and 7.9 FTE on other tasks.
Waste is caused by the fact that the offices of the rheumatologists and residents are approximately 10 minutes walking distance from the outpatient clinic. Sometimes the physician walks 4 times a day from the outpatient clinic to the office and vice versa. Additionally, starting the computers adds up to a significant amount of waste, as it takes a minutes for each computer to start running
CTQ 2 Consultation time
In total, 55 CP consultations with the physician (rheumatologist or resident) were measured. Even though the consultations are scheduled for 15 minutes, the average time of a visit is 20 minutes.
Only 23,6 % of the consultations are completed within the scheduled time and 11% of the consultations take twice as much time as planned (Figure 8).
Descriptive Statistics: totale tijd
Variable N N* Mean SE Mean StDev Minimum Q1 Median Q3 Maximum totale tijd 55 0 19,96 1,01 7,50 8,00 15,00 17,00 24,00 39,00
-22-
Figure 8: Required time of CP consultation. N= 55. 76.4% (42 consultations) of the consultations take longer than the scheduled 15 minutes.
TC
During the consultation sessions the doctors have to perform telephone consultations (TC) as well. Normally, these calls are performed in the last hour of the consultation session. 47 calls were analyzed, out of which 10 were not answered (21,3%).
According to the planners of the TC, they can plan as much TCs as needed, so the number of calls varies significantly. The number of TCs is usually around 4 to 6 per consulting session. But at least once a week the number of calls exceeds 10 (after permission of the physician) for one of the consulting sessions.)
Descriptive Statistics: TC
Variable N N* Mean SE Mean StDev Minimum Median Maximum totaal 31 0 9,581 0,818 4,552 3,000 8,000 25,000 40 35 30 25 20 15 10 20 15 10 5 0 Total time F re q u e n cy 15 3 3 7 9 20 11 2
-23-
Figure 9: Required time of TCs . In total N= 31. 45.2% (14 consultations) take longer than the scheduled 10 minutes
The analysis of 31 calls showed that an average call takes 9,8 minutes (Figure 9). Therefore, not more than 6 calls should be scheduled per hour. However, most of the calls result in more work that needs to be done later in the week: new orders need to be made,
consultation should take place with other colleagues and notes need to be made in the EHR. Most doctors postpone these administrative tasks, and first call all patients instead. One of the reasons why they do this is that the patients are expecting the call. Patients who are not reached the first time are often called for a second time that day or later in the week, outside the consulting sessions. If the call is rescheduled for the next consulting session, an order for a new TC needs to be made. Postponing the administration related to the calls results in extra work outside the consulting sessions, which takes more time than finishing administration directly after the call. There is no data collected about this extra
administration resulting from the calls, but this is estimated to be 5 to 10 minutes per call. When a call takes 15 minutes to finish, including the administrational part, a maximum of four TCs should be scheduled per hour.
24 22 20 18 16 14 12 10 8 6 4 2 9 8 7 6 5 4 3 2 1 0 Total time Fr eq u e n c y 10 1 0 0 2 0 0 2 9 6 7 3 1
-24-
Assuming that the average number of planned calls in the TC session is 8 calls, each taking 15 minutes, including the administration, this will result in one hour extra work per TC session than the scheduled hour for the calls. In total this is 24 hours extra unscheduled work per week for the physicians to complete all TCs.
Total consultation sessions
On average the outpatient clinics (i.e. 1 morning or afternoon of consultation session) finish 34 minutes (N = 24) late. The VRCs are doing better (average clinic finishes 13 minutes late; N=4) compared to the doctors (average clinic finishes 42 minutes late; N = 16). On average physician clinics are overbooked by one to two extra patients and this is one of the reasons why physicians run out of time and clinics finish late
There are eight VRC consulting sessions per week. The average delay of 13 minutes per VRC consultation session results in total of 1.8 hour extra work per week for the VRCs outside the scheduled hours. All other 24 consultation sessions are performed by a physician. This results in total of 16.8 hour extra work (42 minutes per session) a week outside the scheduled hours for the physicians.
CTQ 3 waiting time
Descriptive Statistics: min te laat start consult
Variable N N* Mean SE Mean StDev Minimum Median Maximum min te laat start consul 63 4 7,84 1,30 10,35 -11,00 7,00 36,00
-25-
Figure 10: Waiting time for all scheduled patients. N= 63
In a perfect process patients will not have to wait for their appointment. The data shows that 31.3% of the patients do not have to wait. But on average each patient waits for 8 minutes for his appointment. The agreement at Poli P is that patients will be informed when the physician or VRC run more than 15 minutes behind schedule. However, in this project, waiting time of a maximum of 10 minutes is considered to be acceptable. The data shows that 29.9% of the patients wait longer than 10 minutes. Their waiting time is on average 18 minutes (Figure 10).
Descriptive Statistics: min too late start of consultation
type
Variable Consultat. N N* Mean SE Mean StDev Minimum Median Maximum min te laat start consul Ass 21 1 8,29 2,51 11,49 -11,00 8,00 36,00 Staf 36 2 9,08 1,61 9,68 -6,00 8,00 32,00 VRC 6 1 -1,17 2,46 6,01 -10,00 -1,00 7,00
-26- Updated project objectives and benefits
The data from DMAIC 3 shows the current state of the CTQ (The table 1).
CTQ IST Soll
Outpatient clinic related tasks
68.5% of 11.5 FTE is 7.9 FTE 5.9 FTE
Consultation time 20 minutes 15 minutes
Waiting time 29.9% longer than 10
minutes
0% longer than 10 minutes Table 1: The current CTQ values and the aimed values
The employees work 68.5% (7.9 FTE) of their time on outpatient clinic related tasks. The new aim is to reduce this time with 2 FTE. The average time for a CP consult is 20 minutes and the aim is to reduce the time to the scheduled 15 minutes. The waiting time for 29.9% of the patients is longer than 10 minutes. The aim is to have waiting time no longer than 10 minutes.
DMAIC 4: Identify potential influence factors
Gemba
Observations by the Black Belt and discussions with the employees resulted in a list of 63 potential influence factors. There were 33 general influence factors and 30 influence factors are directly linked to the EHR EPIC (Appendix III list with all potential influence factors that results out of the Gemba walk).
FMEA
A modified Failure Mode and Effect Analysis (FMEA) is used to prioritize all 63 potential influence factors that are the result of the Gemba walk.
-27-
All potential influence factors were rated by frequency, impact and irritation by 14 employees. Some of the potential influence factors are not applicable for everyone, therefore these issues are only rated by a selection of the employees.
A. Improve
‘Performance improvement is the ultimate objective of process managements’ (Slack 2012).
DMAIC 5: Establish the effect of influence factors
CTQ 1 outpatient clinic related tasks
Figure 11: Pie chart of all tasks outside the consultation sessions performed during the DILO measurements
other tasks meetings email and Inbasket interruptions Administration
other outpatient clinic related jobs phone contact waiting time colleague consultation Category 7,1% 2,7% 10,2% 11,0% 20,8% 3,9% 10,3% 15,5% 18,4%
-28-
According to the DILO, 3.7 FTE is used for consulting sessions. 7.9 FTE is used for other tasks, out of which 31.1% is used for administration tasks (20.8% administration and 10.3% email and Inbasket) (Figure 11). Thus, outside the consultation sessions, the employees spend 2.45 FTE on outpatient clinic related administration tasks.
The potential improvement of the most important influence factor of the CTQ outpatient clinic related tasks is 50% of 2.45 FTE which is 1.2 FTE.
The influence factors, related to administration that the Gemba walk revealed, were evaluated by a modified FMEA method. In discussions with the experts the changeability and potential effect of improvement at the CTQ are rated (Table 2: Administration influence factors during other tasks).
Table2: Administration influence factors during other tasks
Most of the identified influence factors on administration have an effect on the CTQ outpatient clinic related tasks as well on the CTQ consultation time. There are two specific influence factors that are not involved in the consultation time. One is the preparing of the
N Extra time Frequenc y Time x Frequenc y Time x Frequency x Annoying Influenced CTQ Changeability Effect B1.Everyone has to find out the details of
EPIC themselfs which makes learning low. There is no coach or follow up workshop to learn new trick in the system
13 3,59 4,68 16,69 52,88 CTQ2 -/+ ++
D7. Preparing the consulting hour takes
twice more time because EPIC is confusing 10 3,44 5,00 15,91 47,36 CTQ2 - ++
D9. To document in EPIC previous texts are
frequently copied into new texts 5 3,50 5,13 15,70 48,70 CTQ2 - +
D14. Orders (no refferals) that are little used
have no clear codes in EPIC 6 4,20 4,40 15,00 62,00 CTQ2 ++ ++
D8. There are many unnecessarily details
that have to be logged into EPIC 8 3,33 4,00 14,50 51,75 CTQ2 - ++
D16. Solving problems in EPIC takes a lot of extra time and often does not results in the right solution
-29-
consultations in the new EHR EPIC. It is estimated that the time needed for consult preparation doubled since the introduction of the new EHR. This includes the time that is needed for answering of InBasket messages.
A second influence factor is problem solving in the EHR. When the employees run into a problem with EPIC, it takes a lot of time to solve the problem. The first solution given by the
EVA team (EPIC helpdesk) often does not solve the problem.
CTQ 2 Consultation time
Figure 12: Average time spent per task during a CP consultation. N=55. Direct patient care: anamnesis + joint score and policy (10.5 minutes). Administration tasks:
preparation, medication, orders and notes (9.4 minutes)
The consultation time is classified into seven task categories. 1. Preparations - the physician prepares the consultation
-30-
2. Anamnesis + joint score - the physician asks about the condition of the patient and checks the joints
3. Medication - the physician checks the medication that is recorded in the EHR of the patient and orders new recipes when needed
4. Policy - the physician discusses with the patient the next steps that will be taken 5. Order for next appointment - the physician makes an order for a new appointment
and a new lab order when needed
6. Other orders - the physician makes other orders in the her, such as x-rays, referrals, additional tests
7. Notes; the physician makes notes into the EHR
The required time for a CP consultation is on average 20 minutes instead of the scheduled 15 minutes. The data showed that administration (orders for appointment, medication, other orders and notes in the EHR, preparing the consultation) during the consultations takes up to 10 minutes, which is 50% of the consultation time ( Figure 12).
The employees spend 31.8% of the 11.5 FTE on consultations. During the consultations, they spend 50% (10 minutes) on administration tasks. Thus 1.8 FTE is spent on administration. Spending 0.9 FTE less on administration is a potential improvement of 50%.
The influence factors during the consultations related to administration, out of the Gemba walk, were evaluated by a modified FMEA method. In discussions with the experts the changeability and potential effect of improvement at the CTQ were rated (Table3: Administration influence factors during consultation).
-31-
Table 3: Administration influence factors during consultation
The conclusion is that the most important influence factors for the EHR are:
- Everyone has to find out the details of EPIC themselves which makes learning slow. There is no coach or follow up workshop to learn new ‘tricks’ in the new EHR system
- Orders that are rarely used have no clear codes in EPIC
- DAS28 form is not properly working and takes more time to complete - There are many unnecessary details that have to be logged into EPIC - The order process for some medications (eg Rituximab) is too complex. The
scheduled 15 min is too short to be able to start a new treatment
- Checking all medication in the system with the patient takes a lot of time
CTQ 3 Waiting time
The data showed that, on average, physicians start the next consultation with a delay of 8-9 minutes. This is explained by the fact that they require more time per consultation than
N Extra time Frequency Time x Frequency Time x Frequency x Annoying Changeability Effect
B1.Everyone has to find out the details of EPIC themselfs which makes learning low. There is no coach or follow up workshop to learn new trick in the system
13 3,59 4,68 16,69 52,88 -/+ ++
D14. Orders (no refferals) that are little
used have no clear codes in EPIC 6 4,20 4,40 15,00 62,00 ++ ++
D4. The DAS28 form is not completed in EPIC, but the results are only listed in the note.
10 3,25 4,44 14,77 55,00 - ++
D8. There are many unnecessary details
that have to be logged into EPIC 8 3,33 4,00 14,50 51,75 - ++
D13.The order proces of some medications (eg Rituximab) is too complex
9 4,43 4,21 14,15 55,78 -- +
D23. It is not possible to see the previous
and new consultation in the same screen 6 3,50 5,20 13,79 30,83 -/+ +
D12. To check all medication during the consultation hour a lot of clicks are necessary which takes extra time
-32-
scheduled. VRCs, on the other hand, start on time, thus there is no waiting time on average (Figure 13) .
Figure 13: Boxplot of waiting time of physicians and VRCs.
The ANOVA test (Appendix IV) showed that the p-value is 0.024, which indicates that the difference in waiting time between the VRCs and physician is significant. The R-Sq is 8.10% which is low, therefore it is not a very important influence factor. The p-value of the equal variance test is 0.338, which in this case means that there is not enough data to prove the difference.
Besides the consultation time, the waiting time is affected by the tasks that the physician or VRC has to finish between consultations.
The waiting time for the VRCs consultation is shorter compared to the waiting time for the consultations with the physician. This is due to the larger fluctuation in required consultation
VRC physician 40 30 20 10 0 -10 w a it in g t im e ( m in u te s)
-33-
time of the patients seen by the physician. The care that is given by the VRCs is less variable than the care that is requested from the physicians. The VRCs tasks are more standardized.
Analyzing the outliers in waiting time in more detail revealed that the big delays, longer than 30 minutes, are all caused by unexpected supervision. Supervision is often planned into the schedules of the rheumatologist, so he or she can supervise the residents and VRCs. The average planned time for supervision is 19 minutes. The data showed that the average time for supervision was 20 minutes. Although the planned time for supervision is sufficient, the planning of supervision is difficult, as supervision is needed at unexpected moments. When a new patient is seen by a resident, there is time calculated for supervision, so the resident can discuss the new patient with a rheumatologist. But during other (control) consultations, patients can also present unexpected symptoms that the resident or VRC needs to discuss with the supervisor.
A rheumatologist that has to give supervision outside the planned time slots and with his own consultation session parallel, will do so and let his/her planned patients wait. Most supervision moments take just a few minutes, but sometimes it can take more than 25 minutes.
The final aim is to have no waiting time for the patients, so the average waiting time should be decreased by 8 minutes.
The influence factors on waiting time are evaluated by a modified FMEA method. In discussions with the experts the changeability and potential effect of improvement at the CTQ are rated (Table 4).
-34- Table 4: Influence factors on waiting time
The conclusion is that the most important influence factors for the waiting time are: - Delayed previous consultation
- Scheduling of patients with the wrong resident due to problems in EPIC the next patient has to wait longer due to extended consultation time of the incorrect scheduled patient
- No clear communication to patients as to when they can expect that the requested prescriptions are send. This can normally take a few days and patients expecting the prescription the next day will call back because they need the prescription right away
- A pager that needs to be answered by physician or VRC - Supervision N Extra time Frequency Time x Frequenc y Time x Frequency x Annoying Influenced CTQ Changea bility Effect C15. Patients are scheduled at the wrong
REST consultation hour 3 4,33 4,50 19,67 78,67 CTQ2 + +
A13. It is for patients not clear (this is not explicitly told to them) that it takes some time before a prescription is send
14 3,27 4,35 14,87 48,31 CTQ3 + ++
A11. Carrying a general department pager (so not a personal pager) during your consultation hour
4 3,38 4,63 15,56 39,88 CTQ3 -/+ +
A6. Residents and staff have no time scheduled during their consultation hour for supervision of control patiënt whom will take extra time
4 3,17 4,00 13,00 29,67 CTQ3 -/+ +
C28. Department Q3 calls for matters that are not urgent and can wait for a respons by email of inbasket
10 3,20 4,00 13,00 34,50 CTQ2 ++ ++
C27. Pager calls are most of the time not urgent, they could for a better moment or the question can be asked by email or inbasket. and how often do you get them)
3 3,00 4,00 12,00 36,00 CTQ2 + +
C18. Sending a prescriptions to a pharmacy or home address takes as much time as asking or explain to the front desk to do this
8 2,75 3,75 11,56 34,09 CTQ3 ++ +
A3. The possibility to perform an ultrasound during the consultation hours adds value but takes extra unscheduled time too
7 3,40 2,90 9,65 22,94 CTQ3 - +
A1. Patients are not well informed when
-35- - Non urgent calls (pager or phone)
- Sending prescriptions by physician or VRC instead of front or back office - Unscheduled ultrasounds
- No information to the patients when the physician or VRC runs behind schedule. Although this is not an influence factor on waiting time, it is an important factor that influences the satisfaction of the patients
DMAIC 6: Design improvement actions
In DMAIC 6 improvement actions are designed for the most important influence factors.
CTQ1 Other outpatient clinic related tasks
The improvement actions for each influence factor in the CTQ other outpatient clinic related tasks are listed in table 5 .
Influence factor Improvement action
Preparing consultations Update of EHR program
Training of employees
Standardize registration in EPIC
Amount of Inbasket messages Protocol/working instructions how to use
the Inbasket function of EPIC
Problem solving EPIC Standardize the process of problem solving
Table 5: Influence factors and improvement action on CTQ other tasks
Preparing consultations
The preparation of the consultations in EPIC takes more time compared to the registration in the old EHR. This is partly explained by the fact that they have to register more details in
EPIC than in the old EHR. An update of the EHR program could help to get rid of unnecessary
registrations and a process that is more aligned with the working process of the employees. When the registration is more standardized, the employees can use more smart phrases and standard text units which are easy to copy into the registration fields.
-36-
Amount of Inbasket messages
The amount of Inbasket messages that the employees have to answer is growing. There are no clear instructions how to use this new communication tool. With clear working
instructions, the amount of non-relevant messages will decrease and it would make it clear what senders can expect when they send an Inbasket message. With clear communication instructions the two other communication tools, email and phone/pager, could be used more effectively as well.
Problem solving EPIC
Reporting EPIC problems to the EVA team (ICT helpdesk for EPIC) is standardized by a standard form on the intranet page. However, a standardized follow up is missing. This results in solutions of the EVA team that are not the answer to the problem.
CTQ 2 Consultation time
The improvement action for each influence factor in the CTQ consultation time are listed in table 6 .
Influence factor Improvement action
Insufficient knowledge about EPIC Training
Unclear codes in EPIC Code list correct codes
Non-functional DAS28 form Update of EHR program
Too many details to register Update of EHR program
Difficult order process medication Update of EHR program
Checking all medication with patient Preparation of the patient before entering the office of the doctor
Table 6: Influence factors and improvement action on CTQ consultation time
The time that is spent on each consultation is longer than the scheduled time. This has, as previously described, a direct influence on the waiting time for patients. But it also affects the workload of all employees. The new EDP resulted in a higher administrative burden
-37-
during the consultations. The performed Lean Six Sigma project gave some directions for improvement such as:
- Insufficient knowledge about EPIC and hick-ups in system in first months of implementation: Training in a more effective and smart use of EPIC. All employees have had a basic training for the use of the EHR program. When the employees started working with EPIC there were no additional trainings. Thus, for the problems the employees encountered when they started using the program, they have to find a solution each for themselves. The problems are often universal and solving them together would save time. It would be useful when all employees will get additional training together. Also, they could teach each other ‘tricks’, so they will learn from each other.
- Unclear codes in EPIC: The problem of unclear order codes in the EHR for referring to other specialties is not only a problem for the KIR department. Other departments at the outpatient clinic have complaints about these codes too. This issue is discussed with the staff of the department poli P. A project group created a list of all common orders in order to make it easier to find the correct order. For all other less frequently used codes in the EHR, the department will make a list of these less used codes. This list will be updated every time someone has a problem finding the correct code.
- Non-functional working DAS28 form and too many details to register:
According to the employees a lot of unnecessary details have to be logged in the EHR and the DAS28 form (a quality score of disease activity of arthritis) is not functional. In 2017 the EPIC team will release an update of EPIC, which needs to include a better functioning DAS28 form. This will save a significant amount of time and will provide data that can be extracted out of the system for other purposes. The DAS28 data will be needed in the near future for care quality systems and can also be used for research purposes.
- Difficult order process medication: An administration process (for ordering medication e.g.) that is more focused on the work flow during the
-38-
- Checking all medication with patients: Better preparation of the patient before entering the office of the doctor or VRC. The patients have to write down when they need new prescriptions and specific questions to the physician or VRC before entering the physician or VRC consultation office. This implies that when a patient has his/her first appointment with the VRC, the consultation with the physician will take less time. The VRC makes sure that the additional tests, like laboratory test, are performed before the appointment with the physician and the joint score is already performed. Some patients do not even need a second consultation with the physician when they are seen by the VRC first.
CTQ 3 Waiting time for the patient
In a perfect process there would not be any waiting time for the patient. The improvement actions for each influence factor in the CTQ waiting time are listed in table 7.
Influence factor Improvement
Delayed consultations Decrease in fluctuation of required
consultation time
Align scheduled time with time that consultation will take
Scheduling the consultation session of the wrong resident
Scheduling which consultation session is planned with which resident
No clear communication to patients about the time prescriptions are sent
Instructions to back office Q3 what to communicate to the patients
A pager that needs to be answered by physician or VRC
No pager during consultation session
Supervision No supervision tasks during own
consultation sessions
Non-urgent calls (pager or phone) Instructions to back office Q3 Sending prescriptions by physician or VRC
instead of front or back office
Instructions to front of back office and standardized process
Unscheduled ultrasounds Planning at the end of the consultation
session or on a special ultrasound consultation session
No information to the patients when the physician or VRC runs behind schedule.
Electronic information about current waiting time in the waiting area and on sms of patient
-39-
Delayed consultations
The main influence factor for the waiting time of the patients is the prolonged consultation time compared to the time that is scheduled. The required care of the physicians is more variable than for the VRCs, but there are still some patient groups that can be treated in a more standardized way. The start of so called zorgpad is a way of the standardized care for specific patient groups.
Alignment of the time needed and planned for consultations can be solved by taking two different directions. Either the planned time is prolonged or the needed time is shortened. Prolonging the consultation time is hard to accomplish. Scheduling more time per patient lowers the production. A lower production results in lower revenues, unless the revenue per consultation increases due to the new EHR that can improve the quality of care. The other way to solve this imbalance regarding the consultation time, is to decrease the time needed for each consultation. See table 7 for options how to improve waiting time.
Scheduling at the consultation session of the wrong resident
Patients were planned at the wrong consultation session with a resident due to the fact that there was unclarity about which resident was doing which consultation session in EPIC. The project group had a close look at the schedule of the consultation sessions of the residents and made a new scheme and the problem was completely solved. There were no wrong bookings after this anymore.
Supervision
A suggested improvement could be that the rheumatologist is not seeing patients simultaneously with supervision duties. At which time a rheumatologist carries out the supervision duties would be a personal choice. Some rheumatologists would perform supervision at a time that they do not have consultation sessions simultaneously. Some of the supervisors are working in their own office which is a 10 minute walk away from the outpatient clinic. When supervision is needed the patient of the resident or VRC has to wait
-40-
around 10 minutes until the supervisor arrives. A supervisor should work in an office at the outpatient clinic instead of his or her own office. The rheumatologist can enter her own digital work station form every computer, so he or she can work on administration tasks or other tasks in the office of the outpatient clinic. A disadvantage for this construction is that the time the rheumatologist has to spend on the outpatient clinic is extended. This would addup to that the already limited time for the rheumatologist to spend on other tasks outside the outpatient clinic. Another pilot could be to change the time schedule of the supervisor when the supervisor is having simultaneously consultations. At present, supervision is scheduled as a 15 minutes break. This can be tested with an experiment in a computer model, something that is out of scope for the Black Belt.
No clear communication to patients about the time it will take to send prescriptions
The back office Q3 already received several times instructions about the handling of
incoming calls from patients. A new phone call routing schedule (see appendix V) is adapted and discussed with the back office Q3. Still, communication is a recurring issue. This is similar for the unclear communication to patients about the time it will take to send the
prescriptions, which results in impatient patients who will call for a second time. So the improvement actions do not only involve instructions of the employees at the back office Q3, but the team has to adapt the process in such a way that the improvement is sustainable as well. This includes training of new employees at the back office Q3. The fact that this issues keeps coming up suggests that a direct feedback loop is required. When the process is not running in accordance with the anticipated perfect process, this should be noticed immediately so action can be taken directly.
A pager that needs to be answered by physician or VRC
The pager that is now carried during the consultations could be handed over to the front desk Q2 employees during the consulting sessions. Thus there will be fewer disturbances during the consultations. The front office Q2 will answer the pager and take the message. Only emergency calls will be transferred into the physicians or VRCs consulting room.
-41-
The task of physicians and VRCs sending prescriptions to the pharmacy themselves should be delegated to the front or back office. A clear standardized working instruction is needed for this.
Unscheduled ultrasounds
The ability to directly perform an unplanned ultrasound to confirm the diagnosis adds high value for the patient. The ultrasound is performed by the physician consulting the patient. Normally, performing an unplanned ultrasound results in extra time with the patient, which again results in a longer waiting time for the next patient. The option to let another
physician perform the ultrasound is suboptimal for care. In some cases it is an option to ask the patient to wait until the end of the consulting hours to perform the ultrasound.
However, when the ultrasound is needed for one of the first patients in line, the patient has to wait multiple hours. A second option is to make a new appointment on the next
ultrasound consulting session preferable with their own physician.
No information to the patients when the physician or VRC runs behind schedule.
Clear communication is essential for patients´ experience. Although it is agreed that the employees inform the front desk Q2 as soon as they run more than 15 minutes behind schedule, in practice this rarely happens. In the ideal world the EHR would be registered automatically when a new patient enters the doctor’s office. The system then communicates this automatically to a screen in the waiting area. Even better, it is communicated to the mobile phone of the next two patients. In that case, the patient is able to use waiting time differently, such as going for a coffee. Until this automatic system is developed the
employees at the front desk Q2 should keep track of the delays of the different consulting sessions. A patient who leaves the doctor’s office normally goes directly to the front desk Q2 to make a new appointment. The front desk employees can check the difference between the time when the patient is at the desk and the end time of the patients’ consultation. This improvement does not directly affect the waiting time, but can make a big difference in the experience of the patient.
-42- B. Control
DMAIC 7: Improve process control
It is important that at the end of the project control systems are created to ensure that the improvements don’t deteriorate over time. Within Lean Six Sigma this can be performed by implementing control points, Poka Yoke (mistake proof), control pyramid, 5s (Sift, sort, shine, standardize, sustain) and visual management.
The control plan defines the process control for the improvements. It specifies what to measure and what is needed for the follow up. The control plan describes how often the data should be analyzed and who is responsible to act on problems (Table 8). The basic statistics can be analyzed every first week of the month for irregularities and trends.
Table 8: Control plan
One of the influence factors on the CTQ 1 outpatient clinic related tasks is the non-urgent calls of the Q3 office. As mentioned earlier, this process is evaluated earlier and still causes problems. Even with a clear communication plan, the problem will reoccur now and then. A direct feedback loop is essential, so there is a direct correction of problems. Daily
registration of non-urgent calls by the physicians and VRCs would be preferable. This gives quantitative results, but it is an extra administration task for them as well. This is something that should be avoided.
More qualitative data could be obtained when the non-urgent calls are discussed every week during the polibespreking. The VRCs are part of the department Poli P and join the weekly stand ups. Therefore, the VRCs can communicate the issues that are discussed at the
Measurement Who How Where When Reporting Norm / spec. Which OCAP
Waiting time Automatically Computer system Computer Every patiënt Electronic board maximum of 10 minutes
Consult time Automatically Computer system Computer Every patiënt Weekly to champion
maximum of 10 minutes delay
Discuss at monthly doktersoverleg
Total consult time Physicians and VRCs Logging Paper form
Start and finish
consulting hour Weekly to champion
max 15 minutes delayed Discuss at monthly doktersoverleg CONTROL PLAN
-43-
polibespreking at the weekly stand up at Poli P. Together with the management of Poli P, the
VRCs can take responsibility to act on problems and instruct new Q3 colleagues when needed.
The CTQ 2 consultation time is not yet controllable by the electronic patient system. The system can log the start and end time of the consultation. However, doctors and nurses do not close the consultation directly after the patient leaves their office, because they enter additional data after the patient is ready. But a switch to another patient file is registered by the system and this can be used as the start and end time of the consultation.
For the CTQ 2 the total time of the consultation session could be logged manually, by note arriving and leaving time of the KIR employees at the outpatient clinic. Automatic
registration would be preferred and the login and log out time at the computer in the
consulting room can be used for this. The physician or VRC notes in a database or on a paper at the front desk Q2 the start and end time of their consulting session. This will be checked by the Champion weekly. When additional information is needed to explain irregularities, the champion will talk to the physician of VRC involved and collect the data about the irregularities. Every month the data can be discussed at the doktersoverleg with all employees. The Champion will be responsible to act on problems.
The CTQ 3 waiting time can be controlled by the electronic patient system. This system logs when a patient is registered at the front desk Q2. It also registers when the doctor or nurse opens the patient file or when the patient makes an appointment for a next visit at the front desk Q2. When this is incorporated in the EHR, the data can be used to detect problems early. These data will be mentioned weekly at the Friday morning polibespreking and when problems occur they will be discussed in detail at the monthly doktersoverleg. In an Out-of-Control Action Plan (OCAP) it is described which interventions are required on frequently occurring problems. The head of the outpatient clinic of the KIR department will be responsible to act on problems and updates the OCAP.
DMAIC 8: Close the project
-44-
When all improvement actions are in place the time spent on outpatient clinic related tasks will decrease from 7.9 to 6.7 FTE. The time spent on CP consultation decreases from 20 to 15 minutes, which will save 0.9 FTE (physician). The time spent on TCs will be reduced from 15 to 7.5 minutes, which will save 0.7 FTE (physician). In total 2.8 FTE of waste time will be saved, which can be spend on other tasks such as research and education.
The waiting time for patients will also improve, so no patients have to wait longer than 10 minutes (Table 9).
CTQ IST Soll Saved FTE
Outpatient clinic related tasks
7.9 FTE 6.7 FTE 1.2 FTE
Consultation time TC 20 minutes 15 minutes 15 minutes 7.5 minutes 0.9 FTE 0.7 FTE
Waiting time 29.9% longer than 10
minutes
0% longer than 10 minutes
Table 9: Potential benefits of improvement
The biggest influence factor for the process is the administration in the EHR. In the perfect EHR all required information is recorded only once. However, the EHR is extensive and the many different tabs make it complex and confusing. The EHR forces to capture more details and includes detailed questionnaires that must be answered. The record information is increasingly expanding. In addition, in the AMC the new EHR EPIC is just running for a year. There are still starting up problems with the new program. The development of the EHR requires extra commitment of staff and the fact that everything is still not working perfectly brings frustrations. This is also evident in this project. Besides ensuring that the problems of the EHR are getting resolved, it also helps to clarify to the employees why they need to capture certain details in the EHR and adopt the program to the wishes of the employees. ? This is motivating and, combined with feedback of data, it reduces the burden experienced by completing administrative tasks.
