Thesis MBA-HC
Optimization of the Traumatic Geriatric Hip Fracture
Care Pathway, by a Lean Six Sigma project.
Jan Rombout 10973133 12-02-2018
I. Executive summary
This Lean Six Sigma (LSS) project was undertaken to shorten Length Of Stay (LOS) of patients that underwent surgery for Traumatic Geriatric Hip Fracture (TGHF). The aim of this LSS project was to improve quality and reduce costs.
At the ADRZ a TGHF ‘Care Pathway’ was implemented in 2016. The trauma surgeons and nurses are since then using a dedicated TGHF Electronic Patient Dossier (EPD). At the start of this project, the LOS was 8.9 days. LOS is defined as admittance time in days, starting from actual admittance for TGHF, ending on discharge. When looking at literature, a mean LOS of 4.8 days is possible, which is achieved at the Bronovo Hospital in The Hague.
To achieve a significant reduction of LOS in the TGHF group, an improvement scheme was developed. This new, improved way of working was introduced by executing an LSS project, which is the subject of this thesis. The DMAIC steps (0-8) were followed, which consist of a Define phase (DMAIC 0), Measure (DMAIC 1-2), Analysis (DMAIC 3-4), Improve (DMAIC 5-6) and a Control phase (DMAIC 7-8).
The only ‘Critical To Quality’ (CTQ) parameter that turned out to be relevant is: mean LOS for TGHF treatments.
The ‘waste’ found in the analysis phase consisted of unnecessary waiting for transfer to a rehabilitation facility after the patient had already recovered from surgery. The aim of the improvement was to reduce waiting time for transfer.
The focus of the project was to speed up announcement for rehabilitation, within 24 hours after surgery, to the proper rehabilitation facility. This LSS project was called ‘Fast Track Transfer’ (FTT). A checklist was introduced to be filled out during the first ward round postoperatively. The Transfer Nurses (TN’s) were informed of a TGHF patient that same morning.
The checklist was aimed to identify the patients that were relatively healthy. They could, theoretically, be announced to the rehabilitation centers the first postoperative day. As patients waited about 22 hours for surgery, the announcement would then be on the second day of hospital stay. Discharge of the patient occurs four days (mean) after the announcement to a rehabilitation facility, which leads to a feasible LOS of 2 + 4 = 6 days.
At the start of the study, in the first cohort, the announcement of the TN’s to a
rehabilitation center of an eligible patient occurred 4.9 days (mean) after admittance, for the TGHF group as a whole. After the checklist was introduced, in the second cohort the announcement was made after 3.7 days (mean) after admittance, for the whole TGHF group. This reduction of 1.2 days, between admittance and announcement for
rehabilitation, is significant. The mean improvement in LOS for TGHF was from 8.9 to 7.8 days. However, this 1.1-day reduction was not significant.
From this study follows that it is possible to improve the administrative processes on the wards. The introduction of the checklist sped up the internal processes of earlier
involvement of the TN’s and earlier announcing the patients to the rehabilitation centers for the group as a whole.
However, changing the medical processes of the surgeons and Physician Assistants (PA’s) proved much more resilient than anticipated. It turned out that the first few days postoperatively the ‘wait and see’ policy was still followed by trauma surgeons and PA’s. Although uncomplicated patients could be identified via the checklist, persistently only the second or third day postoperatively preparations for release were initiated (three or four days after admission), by filling out the right forms. The TN’s had to wait until then, before they could announce the patient to the appropriate rehabilitation center.
In this project, the medical process was resilient to change. Although the formulated improvements were acknowledged in multidisciplinary meetings, the medical
professionals found it hard to comply with the new way of working. The medical staff found it difficult to appoint certain patients for an FTT. 'Change Management’ could provide a solution on how to achieve the desired adaptation of medical routines.
One of the five key points of the Promic 2.0 change model is ‘focus’. Routines can be very deeply automated, to the point that they are ‘unconsciously’ performed. When not properly addressed, i.e. by referring to a ‘critical incident’, old routines stay the same.
Four recommendations for improvement of the TGHF Care Pathway are formulated, for further reduction of LOS.
Firstly, an updated ‘first-postoperative-day’ (two days after admittance) checklist is advised. The most important point is to set an ‘expected discharge date’. When this
Everyone works together to achieve this anticipated date, by ‘pulling’ the patient through the hospital system. This could lead to a significant reduction of TGHF LOS.
Secondly, providing feedback helps to keep the medical professionals involved. A dashboard is needed about the expected discharge date and mean LOS achievements for TGHF patients. It is essential to monitor the percentage adequately estimated expected
discharge date on the first postoperative day, and what the actual mean LOS for TGHF of
the previous month was. By the information from this TGHF dashboard further adaptation of the TGHF Care Pathway, if necessary, can be undertaken.
Thirdly the communication to the rehabilitation centers can be improved. The announcement of an eligible patient can be quicker. The short time aim is the first day postoperatively (second day of admittance), but perhaps the announcement can even be made the first day of admittance and even preoperatively. Also, the medical information can be more accurate. It makes the assessment by the rehabilitation physicians easier and can speed up the application process.
Fourthly, when the right agreements between the hospital and rehabilitation centers are made, the acceptance period could be shortened from four days to two days. The follow-up of this TGHF project aims to reduce LOS further from 7.7 to 5.0 days. Theoretically, LOS could be 4.0 days, but there are patients with complications. The few complicated cases tend to accumulate a lot of hospital days. Therefore a mean LOS of 5.0 days is realistic.
Shortening of LOS for TGHF from 7.7 to 5.0 days will lead to an actual reduction of 2.0 Full-Time-Equivalent (FTE) nursing staff which frees around € 110.000 annually on the budget. When fewer nurses are needed for TGHF patients, one hospital bed can permanently be closed.
A shorter hospital stay is beneficial for the patient, who can start rehabilitation early. Short hospital stay is also cost-effective.
Once this new way of working is implemented, other Care Pathways can be adapted in the same way. Examples are neurologic patients after Cerebro Vascular Accidents (CVA), and lung patients after acute exacerbations and lung infections.
Once the new way of working is useful for quicker transfer of TGHF patients, it can be expanded to other acute ailments and achieve a hospital-wide LOS reduction.
A. List of abbreviations AAW Acute Admittance Ward
ADRZ Admiraal De Ruyter Ziekenhuis (hospital name) CBS Centraal Bureau Statistiek
CTQ Critical To Quality
DMAIC Define, Measure, Analyze, Improve and Control ED Emergency Department
EPD Electronic Patient Dossier
FMEA Failure Modes and Effects Analysis FTT Fast Track Transfer
GTU Geriatric Trauma Unit LOS Length Of Stay LSS Lean Six Sigma
MDO Multi Disciplinair Overleg (multidisciplinary meeting) NH Nursing Home
OR Operation Room PA Physician Assistant
TGHF Traumatic Geriatric Hip Fracture TN Transfer nurse
Table of contents I. Executive summary 1 A. List of abbreviations 4 II. Introduction 6 III. Methods 8 A. Lean 8 B. Six Sigma 9
C. Lean Six Sigma 9
D. Leuven’s 7-steps model for Care Pathways 10
E. Background on TGHF surgery 11
IV. Project 14
A. DMAIC 0: Define the project 14
B. DMAIC 1: Define the CTQ and measurement plan 19 C. DMAIC 2: Validate measurement procedures 20
D. DMAIC 3: Diagnose the current process 21
E. DMAIC 4: Identify potential influence factors 24 F. DMAIC 5: Establish effects of influence factors 27
G. DMAIC 6: Design improvement actions 29
H. DMAIC 7: Improve process control 32
I. DMAIC 8: Close the project 33
J. Results 34
V. Critical reflexion, conclusions and recommendations 37
A. Earlier announcement for rehabilitation 37
B. ‘Change Management’ for medical routines 39 C. TGHF checklist first postoperative ward round 40
D. Keeping control 42
E. Improving communication 43
F. Estimated revenues 44
G. Conclusions about realized goals and follow-up projects 46
VI. References 48
VII. Appendices 51
A. Statistics of the compared cohorts: 1 = Jul-Aug, 2 = Sep-Oct 2017 51 B. Example of relation between FTE nursing staff and LOS reduction 56
C. Mortality 57
D. Use of GTU 58
II. Introduction
Traumatic Geriatric Hip Fracture (TGHF) is a commonly occurring event. TGHF patients are over 70 years of age, have a hip fracture, due to an explaining trauma (fall). At the Admiraal De Ruyter Ziekenhuis (ADRZ) around 250 patients, each year get surgery for TGHF. The ADRZ is a county hospital in the province of Zeeland in The Netherlands and takes care of a population of around 250.000 people, of which 22% is over 65 Years (CBS Statline data).
In hospital care, as in any organization, cost-effectiveness is essential. In the county of Zeeland, a 19% increase of seniors is expected from 2015-2025. They need more healthcare, which is the first reason for increased demand. The second reason for more healthcare demand is that
patients want to expearience personalized, tailored care, which takes more time of the healthcare providers. Rapid developments in healthcare introduce new (and expensive) treatment modalities; these are the third reason. To keep up with increasing demand, healthcare costs somehow have to be restrained (Slenter, Meeren et al. 2015).
The production industry has a long history of process improvements with the aim to make the production process as cost-efficient as possible. A well-known method is the ‘Toyota Production System’ (TPS) that is the basis for ‘Lean’ and also Lean Six Sigma (LSS). These two latter
methods can be applied to healthcare and will be elaborated on in chapter III ‘Methods’.
The basis of process improvements in Lean and LSS is to make work floor professionals aware of their processes and empower them to make improvements. When healthcare professionals have insight into costs of healthcare, i.e. how many hospital days they induce, these professionals can start improving efficiency in a way that aligns with proper, high quality, healthcare (Kaplan and Porter 2011).
A shorter hospital stay after TGHF surgery is seen as better for the patient. Preferably patients are released as soon as possible to home, or else to a rehabilitation facility. Timely discharge can prevent complications such as hospital infections, pressure ulcers and disorientation (Schimmel 2003, García-Alvarez, Al-Ghanem et al. 2010).
Nikkel discovered a shorter hospital stay was correlated with lower 30-day mortality. But shorter hospital stay could also be attributed to lower co-morbidity. Concluding that shortening hospital stay lowers mortality is not defendable. However, It can be concluded from Nikkel’s research, that shorter hospital stay does not influence 30-day morbidity negatively (Nikkel, Kates et al. 2015).
Quicker release after TGHF stands for better quality and leads to faster rehabilitation and secondary to fewer costs (Lawrence, White et al. 2005).
In healthcare, a lot can be gained by improving the efficiency of medical processes. Although many physicians and managers remain resistant to quality improvement methods, that were created for manufacturing, more and more healthcare organizations are adopting process
improvement methods, such as the ‘Six Sigma’ principles, achieving higher quality and better cost-efficiency (Sherman 2006). The knowledge acquired in the industry can be translated to healthcare by looking for the similarities and copy the proven best practices for these topics (Mast, Does et al. 2012).
Applying Six Sigma projects to healthcare processes leads to cost efficiency. Vd Heuvel has proven that Six Sigma projects are feasible to reduce LOS (Heuvel, Does et al. 2005, Heuvel 2006, Koning, Verver et al. 2006).
Treatment for a disease, such as surgery and postoperative care for TGHF, is a fairly linear and predictable process. Treatment processes can benefit from an LSS approach because this part of medicine closely resembles a production process.
From the literature, it is known that it is possible to improve the quality of care and at the same time to achieve a more efficient allocation of resources and hence increase the hospital’s
efficiency (Schoonhoven and Brilleman 2011).
This LSS THGF total Length Of Stay (LOS) reduction-project aims to combine quality
improvement on the one hand, and at the same time cost reduction on the other hand. To be able to achieve the best improvement possible, all stakeholders in this process were involved. At the ADRZ the LOS for TGHF was at the beginning of the project mean 8.9 days. Although this figure is within the range as mentioned in the literature (Niemeijer, Flikweert et al. 2013), we still felt it was possible to reduce LOS. This LSS project was aimed at reducing LOS for TGHF patients, by speeding up ADRZ care processes.
III. Methods
A. Lean
The customer is solidly placed at the center of ‘Lean thinking’. What the customer values is taken as starting and endpoint for every process improvement. Lean identifies the ‘value’ that is added for the customer. When analyzing processes with the aim to increase efficiency, a hunt for ‘waste’ is undertaken. Waste does not add value for the customer and is thus best eliminated.
Lean production evolved from the ‘Toyota Production System’. Eiji Toyoda and Taiichi Ohno are seen as its founding fathers. Womack describes the rise of Lean production in his book ‘The machine that changed the world: the story of Lean production. (Womack, Jones et al. 1991).
In Japanese terms, three forms of waste are distinguished: “muda, muri, mura”. Muda stands for non-value adding activities: a.k.a. waste. Muri means overuse of resources, such as staff, equipment or digital systems. Mura is the equivalent of unwanted variation. Kruskal describes these three forms of waste in healthcare (Kruskal, Reedy et al. 2012). Distinguishing between value adding activities and waste is not always apparent (Young and McClean 2008). There are a few ways to gain insight into processes and how to remove waste. Undertaking a ‘Kaizen’ can induce a small, incremental improvement. Black has described this in his book how Lean can increase efficiency in healthcare (Black, Miller et al. 2016).
Krafcik describes that Lean factories are much more able to produce a variety of products/ models on the same production line and in the same factory (Krafcik 1988). Healthcare professionals regard every patient as ‘unique’. Lean enables on the one hand to reduce waste, and on the other hand, make adaptations to individual customers demands. This makes Lean applicable to healthcare.
In Lean organizations, trust exists between employees and managers. Managers know they can’t grasp the detailed knowledge and experience of their subordinates and trust that they are working as efficient as possible. In healthcare, the highest educated (and paid) workers are on the work floor, I.e., doctors en specialized nurses. Because Lean thrives on participation from the work floor, it is adaptable to healthcare: medical professionals can funnel their experience into improvement projects.
To achieve efficiency improvements, Kaizens are undertaken to study and improve (parts of) a process. Virginia Mason Medical Center installed a ‘Kaizen Promotion Office’.
The employees trust their management to facilitate them in their work and support their development of improvements, that result from i.e. kaizen activities (Kenney 2011).
B. Six Sigma
At the heart of Six Sigma lies the goal of achieving the highest possible quality. Six Sigma strives to reduce ‘costs of poor quality’ and is, like Lean, customer-centered. By
standardization and uniforming work processes, defects are reduced to the lowest level possible. This higher standard has a direct, internal, effect, i.e. by reducing ‘rework’, and thus improving efficiency. There is also an external effect: products are better quality and more reliable than that of competitors, which leads to better market standing and
premium prices (Bisgaard and Freiesleben 2004).
The history of Six Sigma begins at Motorola Corporation, where Bill Smith introduced this system. It was then further developed by Jack Welch at General Electric. Six Sigma has an emphasis on analyzing the process by measuring. The results are then fed back to the working floor in comprehensive overviews (Harry and Schroeder 2001).
The Six Sigma ‘Black Belt’ is in charge of big projects, and ‘Green Belts’ run smaller projects. At General Electric being a Black Belt was a full-time job, but temporarily (Mast, Does et al. 2012). The Black Belt plays a role in analyzing the process, improve the
quality, together with the employees on the work floor. Bringing the process under
statistical control is one of the critical points. Statistics are not for everyone. Therefore the work floor employees that participate have to rely on the Black Belt for quality control processes (Kooy 2002).
Six Sigma introduced the DMAIC; Define, Measure, Analyze, Improve and Control. This very structured approach is also applicable to healthcare (Taner, Sezen et al. 2007). Taner also describes six attributes of healthcare: Safe, effective, patient-centered, timely,
efficient, and equitable. He advised work-floor professionals to ask senior management for guidance for training and help with the implementation of the improvements.
C. Lean Six Sigma
LSS is a combination of Lean and Six Sigma. According to George, this combination enables customer satisfaction by process improvements and at the same time bring a process under statistical control. Both enable the reduction of the cost of complexity (George 2003).
LSS strikes a balance between the need for standardization, which reduces rework, and individualized ‘products’, or ‘treatments’ in healthcare. There is a fine balance
The work floor employees have to give their input on how they want to achieve the quality
improvement and waste reduction. They are ‘empowered’ by LSS. Management, aided by
the Black/ Green Belts, places possible improvements in context and takes care of ‘office politics’ (i.e. dealing with other departments). Management, together with their
employees, get ‘in control’ of the processes and can continuously strive for optimization. D. Leuven’s 7-steps model for Care Pathways
To integrate evidence-based medicine into daily practice the 7-steps Care Pathway model was introduced (Aeyels, Veken et al. 2016). A multidisciplinary team of healthcare
professionals can use the seven consecutive steps to come to a sound description of a
Care Pathway. It is possible that after step 1 (screening) is decided that the chosen topic
is not suitable for a Care Pathway (Table 1).
The 7-steps model lacks the precision of LSS, but it gives structure to multidisciplinary meetings. Its structure resembles very much how healthcare professionals work. This method also has a clear view of what ‘added value’ could mean for a patient.
Professionals are stimulated to involve patients. A few LSS tools also appear in this method, like the Failure Modes and Effects Analysis (FMEA).
The Care Pathways-methodology is very apt to improve contact between different groups of healthcare professionals, i.e. from different departments that all are involved in the same care process. Also, patients can be structurally involved when designing and doing revisions of a Care Pathway. Transfer moments can be addressed appropriately because all people involved participate in the development and attend the
multidisciplinary meetings (Hoeve and Vries 2014). This method focusses on the ‘soft parameters’ rather than ‘hard data’.
Care pathways can lead to a significantly lower length of hospital stay (Vanhaecht 2007). It is, however, a complex intervention which still has to be further developed.
Table 1: Overview of Leuven’s 7-steps 1. Screening
2. Project management 3. Research and fact-finding 4. Development
5. Implementation 6. Evaluation
A multidisciplinary team of clinicians, healthcare managers and patients is formed to develop a Care Pathway. A Care pathway needs continuous follow up to keep improving and to keep up with new developments in healthcare.
A Cochrane review form Rotter states: “Clinical pathways are associated with reduced in-hospital complications and improved documentation without negatively impacting on length of stay and hospital costs.” (Rotter, Kinsman et al. 2010). Vanhaecht comments on this Cochrane report that Care Pathways are very complex and every organization has its way of installing a multidisciplinary team and implementing the changes derived from the development of a Care Pathway, which makes Care Pathways difficult to compare
between institutions (Vanhaecht, Ovretveit et al. 2012). The same follows from a literature study on Care Pathways on Hip fractures (Leigheb, Vanhaecht et al. 2012). There are indications that a Hip Fracture Care Pathway is beneficial to the patient, but hard evidence is not apparent.
Despite the lack of precision, it very closely connects with the way healthcare
professionals think. LSS is still a tool developed for business. When LSS is combined with key-elements of the 7-steps model, it can be much more suitable for healthcare. Also, the 7-steps model could benefit from better measurements and more statistics, as promoted in LSS.
E. Background on TGHF surgery
In a comprehensive review in the British Medical Journal the key indicators are listed (Parker and Johansen 2006). The basis of the following text is derived from this overview; local peculiarities are added.
Hip fracture, especially in the elderly, is relatively common. It is the most occurring indication for acute orthopedic surgery. At the ADRZ the team of trauma surgeons consists of trauma surgeons and trauma orthopedics.
Almost all hip fractures are operated, conservative treatment is not sufficient and very painful, which leads to permanent disablement. Sometimes it is feasible when the patient is obvious moribund.
In elderly patients with intracapsular fractures (Figure I) either only the femoral head is replaced, or the total hip is replaced (Figure II). When the fracture is extracapsular, the most used device is the sliding hip screw (Figures III and IV). Most patients can put their full weight on the operated hip, straight after surgery.
Mortality is 5-10% after 30 days and about 30% after one year. Multidisciplinary rehabilitation is needed before the patient can go home.
Surgical complications account for 27% of postoperative hospital days (Foss, Kristensen et al. 2006). Postoperative complications are urinary tract infections,
pneumonia, superficial wound infections, deep wound infections. “Transfusion and longer waiting time for surgery have been associated with the septic complications in elderly patients treated surgically for hip fracture” (García-Alvarez, Al-Ghanem et al. 2010).
Figure I: possible areas for a hip fracture (Parker and Johansen 2006)
Figure II: Radiogram of a Thompson hip hemiarthroplasty
(Pototschnik 2017, TDM 2017)
Figure IV: Dynamic Hip Screw (TDM 2017) Figure III: Intramedullary hip
IV. Project
A. DMAIC 0: Define the project
In the first phase, the project is defined. Included in this project are patients that have a Traumatic Hip Fracture with an explaining event, are Geriatric (age over 70) and
underwent surgery. The aim was to speed up the hospital processes, aimed at earlier discharge. A summary is given in Table 2.
1. SIPOC
To describe the process, a SIPOC overview was designed (Table 3). SIPOC stands for Supplier, Input, Process, Output, and Customer.
The suppliers of this process are the ambulance service and General Practitioners (GP’s). Input consists of patients with TGHF. The process covers surgery and
postoperative care. The output is patients with a treated TGHF. The customer is in the first place the patient. But also the rehabilitation centers and the insurance companies can be seen as customers.
When studying the TGHF patients, three categories could be discerned. 1. Patients who can eventually go back home, mostly after rehabilitation.
2. Patients who cannot go back home, but need a place in a Nursing Home (NH). 3. Patients who already live in an NH can be discharged the second or third day
postoperatively, back to their nursing home.
The first group can benefit most from a change in the medical process, because they are eligible for transfer the third day postoperatively, but are now transferred the seventh day postoperatively.
The rehabilitation centers need around four days to anticipate before a transfer is made. After the TN’s announcement, the third day postoperatively, that a patient is ready for transfer, it then takes another four days until the patient is transferred. At least the last two days of the mean hospital stay of 7.7 days are just waiting days.
2. Benefit analysis
In the first half of 2017, the mean LOS was 7.7 days. At the time of starting this project, LOS had increased to 8.9 days. The moment of notification to the rehabilitation center (if applicable) was estimated around the fourth day of admission.
Quicker discharge leads to fewer complications, such as urinary tract infections, pressure ulcers and disorientation (Schimmel 2003, García-Alvarez, Al-Ghanem et al.
2010). Applying the appropriate level of care, i.e. by inducing quick mobilization, delivers a better quality of care (Nijmeijer, Folbert et al. 2016). Better quality of care processes reduces, generally speaking, costs in healthcare (Schoonhoven and Brilleman 2011). Timely discharge to a rehabilitation facility after TGHF enhances quick recovery and reduces costs of hospital stay at the same time. Less use of hospital beds leads to less need for nursing staff.
The financial aim of this project was to achieve a reduction in costs of 2 FTE ward nurses. To achieve this reduction, a mean shortening of LOS with 1.8 days is necessary. Berden calculated, in an example, that one hospital day reduction leads to a reduction of 0.0045 FTE nursing staff. This example was comparable to the ADRZ wards (Berden, Berrevoets et al. 2016). 247 * 1.8 days * 0.0045 = 2.0 FTE nursing staff. The aim of this project is to reduce the mean LOS from 7.7 to 5.9. The precise calculations are given in Appendix B.
There are also soft benefits from improving the internal care processes around TGHF. Patients get better explanations as what to expect from their hospital stay. The process is much clearer for doctors and nurses, because of the checklists. This leads to more
efficiency during ward visits and better communication.
Strategic benefits are also to be expected. Once the internal processes are robust and dependable, the ADRZ can also be a reliable partner for the rehabilitation centers and NH’s. Their comment on the current process is that the provided information is
sometimes not very accurate (see next paragraph ‘Stakeholder analysis’). When the internal processes are transparent and the right information is presented, the external partners can hook up their processes to the ADRZ processes. This could improve the patient flow between these institutions.
3. Stakeholder analysis
A stakeholder analysis was undertaken of all medical professionals involved in the TGHF process. In Table 4 the most important professionals are listed. Most of the professionals are categorized ‘good as is’ (green). Only the status of the Emergency Department (ED) nurses and Trauma Surgeons was unclear at the beginning of the project. Both these groups of professionals did not seem to be able to explore their part in the treatment of TGHF patients; they felt ‘good as is’ themselves.
The rehabilitation facilities proved to be skeptical about the project; they did not trust the information from the hospital. This distrust has to do with some unfortunate behavior
in the recent past. Patients were ‘downscaled’ to speed up the transfer process. So patients that were transferred for simple rehabilitation needed much more care than anticipated, which is, of course, more expensive. Rehabilitation centers are ‘not amused’ by downscaling and now perform a thorough assessment of the eligible patients which, of course, takes time.
The table is presented in two parts for readability.
The stakeholders were represented in a two-by-two table to find out where the focus had to be to change the process and speed up discharge from the hospital. From Table 5
Table 4: stakeholder analysis. The most important groups of involved medical professionals are listed, and their stake is described (copied from the LSS charter).
followed that the Rehabilitation facilities and Trauma Surgeons needed the most attention in changing the process. The trauma surgeons were actively involved by the project leader attending their meetings and providing information. The Rehabilitation facilities were informed and gave their formal consent to start an ‘early warning system’ for relatively healthy patients with TGHF.
Table 5: table of influence of the stakeholders
Low Influence High Influence
Positive stake Other involved specialties Team Leader GTU
TN, PA’s, Team Leader AAW Department management Negative stake ED nurses Rehabilitation facilities
B. DMAIC 1: Define the CTQ and measurement plan
The most relevant parameter in this study is LOS. The interventions undertaken should reduce LOS. As this study aims to speed up the application for a rehabilitation place, the time between admittance and notification to the rehabilitation center is relevant: this should be reduced. The CTQ flow-down in Table 6 shows that LOS is directly coupled to cost when the number of beds can be reduced following the LOS reduction.
In Table 7 the two relevant parameters are presented: days until announcement for rehabilitation and LOS in days.
LOS is derived from the hospital patient database, or Electronic Patient Dossier (EPD). Time of admittance is registered at the Emergency Department (ED) when the transfer of a patient to the ED staff is performed. Date of discharge is also derived from the hospital database. The date of announcement to rehabilitation centers was collected from an Excel file, administered by the TN’s.
Table 6: CTQ (Critical To Quality) flow-down
Table 7: Relevant parameters
CTQ Operational definition
Announcement of patient eligible for rehabilitation
Number of days between admittance and notification of the rehabilitation center
Length Of Stay (LOS) Number of days between admittance and
C. DMAIC 2: Validate measurement procedures
Most data are derived from the main hospital computer system. A check on the validity of the arrival time was performed at the ED. It was checked if the date and time registered in the hospital computer were in accordance with the actual arriving time and briefing of the hospital staff by the emergency nurse of the ambulance service.
In three days 30 patients were observed. The observed briefing time deferred from the registered time, with a minimum of -1 minute (early registration) and a maximum of 10 minutes (late registration). The validation actions are summarized in Table 8. The date and time of admission are valid enough to take as a starting point.
The notifications of the rehabilitation centers were retrieved from a separate Excel sheet, administered by the TN’s. These data points were cross-checked with the nursing notes in the EPD. They all checked out: no dissimilarities were found. Because the exact time of rehabilitation announcement was not administered, announcement time was set at noon. Discharge date was checked for all patients with the medical and nursing notes in the EPD; the date always proved correct. However, a reliable time of discharge could not be estimated. Checking on the ward showed that there were two release times: 10:00 in the morning and 14:00 in the afternoon. The actual time the patient physically left the ward was not adequately registered. Often the ward nurse put a convenient time in the dossier. The nurse handed over all paperwork to the patient, and then the patient left sometime later unobserved, with family or by an ambulance. The maximum deviation of release time during observation was 1 hour and 45 minutes. Therefore the release time was set for all patients at noon, which is the best approximation possible.
All dates were retrieved from the digital hospital systems and patient dossiers. Table 8: validation of measurements.
Measurements Validation
Admittance time (start of LOS) Check on ED 3 times ten consecutive patients if date/ time of arrival is correct
Announcement to rehabilitation center Cross check if the TN data are aligned with the medical dossier and nursing notes of 2 months = all TGHF patients in July and August 2017
Discharge date/ time (end of LOS) Check on ward three times three patient if the date and time of discharge are correct
D. DMAIC 3: Diagnose the current process
For all statistical analysis, the Minitab software was used.
A control chart was constructed (figure V) to see whether the process was in control.
o This control chart shows the successive patients from January 1st, 2016 – June 30th,
2017.
o The outliers are all patients with co-morbidity, which causes delay. o Mean LOS is 7,7 days.
When the project started, additive information was collected. In the past, the exact day of notification of the rehabilitation facility was not always documented. Also, the exact date when the TN was involved was unclear from the records. Therefore the two months before the process change was introduced were taken to establish the date of
involvement of the TN and the date of notification to a rehabilitation center. However, the LOS also lengthened in these two months as compared to the previous half year.
Figure V: control chart of LOS, 1,5 years preceding the current project
Table 9: waiting time for notification of eligibility for transfer and LOS during the initiation time of the current project
Mean waiting for notification
In days Mean LOS In days July-August 2017 4.9 8.9
In this cohort, after 4.9 days mean, the patient was announced to a feasible rehabilitation center. Typical waiting time, after the announcement of an eligible TGHF patient for rehabilitation, is around four days before actual transfer. LOS appeared 7.7 days in 2016 and the first half of 2017, and 8.9 in July - August of 2017 (Table 9).
To establish where process improvements can be made a 'value stream map’ was constructed (figure VI). The basis is a flow-chart in which all relevant sub-processes are listed. Unnecessary waiting was the most occurring form of waste (see chapter III: Lean) in the TGHF process.
It turned out that TGHF patients had to wait before the TN showed up at their bedside, to start the transfer process. The TN process took about 2 1/2 days to complete, and then the announcement to a rehabilitation facility is made. Then patients wait for placement at a rehabilitation facility for about four days.
In figure VI the value stream map for the TGHF surgical process is presented. The first part, figure VI.a shows the first phase of work-up and surgery. The second phase, in figure VI.b, shows the postoperative nursing. This map shows at three moments unnecessary long waiting, which are three forms of waste, could be eliminated.
Firstly, patients wait a day for the TN to show up at their bedside. Secondly, it takes two further days before the announcement to the rehabilitation centers is made, because of a ‘wait and see’ policy, to determine if recovery is uneventful. Thirdly, the rehabilitation centers need about four days to prepare for the actual transfer.
Figure VI.a: Value stream map of TGHF patients, treated by surgery. First part: work-up and surgery; remains unchanged (copied from the LSS charter).
See fig. VI.b
= process inefficiencies, in this case unnecessary waiting, which counts as waste. Figure VI.b: Value stream map of TGHF patients, treated by surgery. Second part:
E. DMAIC 4: Identify potential influence factors
This chapter describes the different actions that were undertaken to identify potential influence factors. The information was retrieved from the involved medical professionals. Several brainstorm sessions were conducted, with the different groups separately
(surgeons, nurses, TN). One multidisciplinary session was held. Individual stakeholder interviews were held during the study.
o Brainstorm session/ group meeting: work process
• Trauma surgeons tend to ‘wait and see’ what happens with the recovery of the patient postoperatively. After 2-3 days they ‘release’ the patient and expect the transfer to a rehabilitation center within 24 hours.
• TN’s have to speak to the family to confirm the choice of a rehabilitation center. It takes days to arrange a meeting with the family (mostly a daughter).
• Much paperwork is involved in the transfer process. Filling out forms properly, with the right patient data, is left for last. Many times the TN has to undertake a ‘chase for information’.
• It takes 4.9 days before the TN can send out an announcement to the proper rehabilitation facility. After the announcement, it takes about four days until the patient can be transferred.
In conclusion: the work-up for transfer is not handled the day after surgery but takes many days.
o Brainstorm session/ group meeting: organizational issues
• TN’s need help from the surgeons and ward nurses» Surgeons and nurses need to inform the patient about rehabilitation, and that quick transfer is imminent.
» Surgeons and PA’s should fill out the medical forms timely, correctly, and without downgrading the scores. Downgrading occurs because surgeons believe patients are then transferred quicker. However, downgrading is always discovered and induces distrust at the rehabilitation centers. They then take extra time for a thorough evaluation of the next patient.
• Receiving rehabilitation centers are a tremendous external influencer of the waiting time. Relationship building by being a reliable partner, and providing timely and adequate information, could help to speed up the transfer.
The conclusion is that the whole medical team (nurses/ doctors/ TN’s and
physiotherapists) need to work together and improve the quality of their organization processes, such as the paperwork and communication.
o Expert interviews
• Surgeons indicate they can not take care of poly-pharmacy prescribed by other specialties. Often the medication needs revision, help is needed from internal medicine.
• TN’s tell that the medical information on the transfer forms is sometimes unreliable, and late. This gives rise to confusion, distrust, and delays.
The conclusion is that the transfer of medical information from the hospital torehabilitation centers could be improved. Internal medicine and geriatricians need to be consulted. This is a joined responsibility of surgeons, (transfer) nurses and PA’s.
When studying potential influence factors, an analysis is made of ‘the Best Of the Best’ and ‘Worst Of the Worst’. These are presented here:
o Best Of the Best
• Patients from NH’s get released after a LOS of 3.4 days. They are quickly transferred back to their NH bed (Appendix A; cohort 1).
• Patients that can afford a ‘care hotel’ can also be released 2-3 days post-operatively.
• Hospital Bronovo in The Hague has a contract with an NH that all patients are transferred two days postoperatively; their mean total LOS is 4.8 days (Wansink 2014).
o Worst Of the Worst
• Patients that turn out to have psychogeriatric ailments cannot go to a
rehabilitation center; they have to wait for an NH bed, waiting can take weeks.
• Without a proper assessment by internal medicine, a psychiatrist or geriatrician,NH will not accept the application, which induces delays.
• The family can ask for a specific high-valued NH, which can give rise to further delays.
An analysis of FMEA was undertaken (Table 10). This table showed that the role of the TN’s is vital.
Other parts of the medical process preceding surgery were examined. The workup at ED, waiting time for surgery and surgery duration do not give a very significant delay,
estimated at most 0.5 day. These processes are difficult to assess and change, therefore out of scope.
Three process Inefficiencies were identified, which all consisted of unnecessary long waiting. The first Inefficiency occurred before the first contact between patient and TN. The TN could show up earlier at the patient’s bedside. The second inefficiency occurred while waiting for the announcement for rehabilitation; this process could be sped up. The third inefficiency consisted of waiting for actual transfer, while the patient didn’t need hospital care anymore, which is due to external factors. These three inefficiencies are shown in the value stream map (figure VI.b)
Table 10: Failure Modes and Effects Analysis (FMEA) table for TGHF (copied from the LSS charter)
F. DMAIC 5: Establish effects of influence factors The most important influence factors are
o Time to notify the rehabilitation center.
o If the patient is directly postoperatively at the Geriatric Trauma Unit (GTU). At the GTU
THGF patient are clustered, the know-how of this specific group is better, and a better secondary nursing environment exists.
o Waiting time between the announcement of the patient and the actual transfer to a
rehabilitation center.
There are several recommendations to make improvements.
o The TN should engage with the patient as soon as possible.
• The family can induce delays of several days when not adequately informed by the surgeon of the importance of quick rehabilitation.
• After the announcement, the rehabilitation centers need about four days to anticipate and find an empty bed. At least two of these days are unnecessary. Better communication could speed up the external process.
o Patients preferably stay at the GTU, where secondary nursing conditions are better.
• GTU has a living room; patients are mobilizing better and keep their regular day-night rhythm, which reduces delirium. Deranged patients are not accepted for transfer, which can induce delays of several days.
• PA’s have their office at GTU, which is more efficient than working on other wards. I.e., consultation of other specialties is better guarded.
o Time to notify a rehabilitation center is a parameter that takes an effort to change.
• The medical professionals are wary of postoperative complications. When this occurs, the transfer is delayed and the whole process has to start all over again. However, the chances of complications are low, so only a few patients get delayed after their initial early announcement.
• Also, when the trauma surgeons make a prediction when the patient could be discharged, the TN’s don’t have to wait two to three days postoperatively, before admitting the patients for transfer to an external facility.
It is estimated that when the patients with low co-morbidity are referred within 24 hours postoperatively, the mean total LOS of this group can be reduced from 7.7 to under 6.0 days. The mean waiting time for surgery is 22 hours. The TN announce ideally within 24 hours after surgery the patient to a feasible rehabilitation center. It takes then another
four days before the patient leaves the hospital. This calculates to a LOS of 1 + 1 + 4 = 6 days under current circumstances.
Being at the GTU, which is the best ward to be, helps to achieve a mean LOS of 6 days. Complicated patients (who gather a lot of hospital days) get to be transferred earlier when complications such as infections and delirium are quickly recognized and
adequately taken care of, which is easier at a dedicated ward for TGHF patients. Use of the living room is essential in keeping the patients in good (mental) shape.
But six days is not the absolute minimum. The best case scenario is that all patients are transferred within three days postoperatively.
o When the waiting time after the announcement for actual transfer is reduced to 48 hours, most patients will be discharged around four days. The few complicated cases tend to induce more hospital days. Therefore a mean LOS of 5.0 is more realistic than 4.0 days. The follow-up project aims to reduce TGHF LOS further, from 7.7 to 5.0 days.
o The Bronovo hospital in The Hague has an agreement with a dedicated care home;they can transfer most patients the 3rd day. Key is that the Nebo Rehabilitation center does not select TGHF patients, and the trauma surgeon does ward rounds at this rehabilitation ward. The postoperative medical care is thus the same high level at the hospital ward and the rehabilitation ward at the Nebo care organization. Bronovo has a LOS of 4.8 days (Wansink 2014).
o A LOS of 5.0 days (2.7-day reduction) could result in a reduction of 3.0 FTE nursing staff. As the remaining patients need more care, to stay on the safe side, a reduction of 2.0 FTE is more realistic. This accounts to a reduction of 2 * € 55.000 = € 110,000 (See Appendix B).
G. DMAIC 6: Design improvement actions
When everything runs ideally, and there are no medical or other complications, the following applies to TGHF patients at the ADRZ:
o It is possible to announce within 24 hours after surgery that a TGHF patient is eligible
for rehabilitation, when otherwise healthy (figure VII, value stream map of the desired process).
o 72 hours after successful surgery the patient can be transferred to a rehabilitation
center. Theoretically, all otherwise ‘healthy’ patients are starting rehabilitation when LOS is 4.0 days, at a rehabilitation facility outside the hospital.
An earlier announcement to rehabilitation facilities that they can await a patient with TGHF is possible.
o When the patient is ‘healthy’ (well-controlled co-morbidity) there is no need ‘to wait
and see’; the notification for a rehabilitation place can be given within the first 24 hours postoperatively.
o Early notification should shorten mean LOS within a few weeks after changing the
procedure.
o As a first step a ‘Fast Transfer Track’ (FTT) was designed, eligible patients are included,
and quick notification is induced.
Figure VII: Value stream map of TGHF patients, treated by surgery. Second part; postoperative care: desired situation.
Early notification for rehabilitation for ‘healthy’ TGHF cases, should give a reduction in LOS for the TGHF group as a whole.
A checklist was designed, which enabled professionals to establish if a patient is eligible for FTT. This FTT-form enabled to forecast if the patient can go home after rehabilitation, which is a prerequisite for FTT.
For all TGHF patients, the FTT form is filled out. If they are eligible for FTT, also the medical referral form for the rehabilitation facility is filled out the first morning after surgery. The TN is then notified the same morning, so she can contact the family the same day, to arrange for quick transfer.
All rehabilitation facilities were informed and consented to the change in procedure. TN’s were allowed to announce patients in an early postoperative phase.
The FTT-form was introduced 1st September.
o The Project leader did the introduction of the trauma surgeons, ward nurses, and secretaries.
o The PA is responsible for the correct handling of the forms during rounds on the wards.
o One TN is responsible for the quick reaction from her department.
Endorsing of TGHF patients staying at the GTU was done on multiple occasions.
o Awareness by education of trauma surgeons, PA’s, and ward nurses, that TGHFpatients should preferably stay at the GTU. The project leader makes weekly visits to stakeholders and reminds them. Started September 1st.
o Stimulating use of the GTU living room by the patients, by making it more comfortable, i.e. by providing proper furniture, television, new magazines, and newspapers. Started October 1st.
o Deploying volunteers in the living room to entertain the patients (i.e. card games) and having lunch together, sitting at a table. Extra mobility and sitting (instead of lying in bed) prevents ulcers, lessens the chance on unitary tract infections, keeps patients in their regular day-night rhythm (Schimmel 2003, García-Alvarez, Al-Ghanem et al. 2010). Started November 1st.
The external partners, such as rehabilitation physicians and facilities have to be involved in the medical process as early as possible. It is essential to understand the need for information and provide accurate, adequate and timely medical information.
The first step is to speed up and make the internal ADRZ hospital processes
qualitatively better. Then the announcement of eligible TGHF patients for rehabilitation to the external partners can be made around the second day after admission.
The second step is to involve the rehabilitation centers, so that they can accommodate for the TGHF patients after 48 hours, instead of after four days (96 hours). To make a quick assessment the external partners need adequate and timely information. Direct contact between the trauma surgeon and receiving rehabilitation physician can help in this respect.
H. DMAIC 7: Improve process control
As of November 30th, a new way of working was implemented to improve process control, which consists of the following short time actions:
o Improve feedback about LOS by a dashboard that provides mean TGHF LOS. The trauma surgeon who owns the TGHF Care Pathway is responsible for the continuation of ‘LOS reduction’ (Table 11). He discusses LOS monthly with management and the TN’s. An action is undertaken if necessary (i.e. rehabilitation centers get involved).
o The PA is responsible for discussing the mean TGHF LOS during the MDO(multidisciplinary meeting) at the GTU. It is weekly discussed if the mean six-day goal is achieved (preferably less).
o The short time aim is a mean hospital stay of 6.0 days, which is aimed to be reduced in the longer run to 5.0 days.
Further improvement of feedback on the transfer process is achieved when: o The percentage of adequately formulated expected release date the first day
postoperative is available from a dashboard and is regularly discussed.
o The waiting time until notification is reported and discussed by TN and management on a monthly basis.
o It is documented when hospital treatment is completed, staying in the hospital is no longer necessary, and the patient is only waiting for transfer. This moment lies after the announcement to a rehabilitation center, around four days after admission. The waiting time between this ‘end of treatment’ moment and actual transfer is the basis for
discussions about improvements between the trauma surgeon and rehabilitation centers.
I. DMAIC 8: Close the project
31st October: first evaluation, with senior management and the leading Trauma Surgeon. The project is finalized, and the end product will be delivered to the trauma surgeons, TN’s, surgical nursing wards, and their management. The aim is to develop a LOS dashboard, which will be discussed monthly in the different meeting of the involved healthcare professionals (doctors, nurses, and management). On the17th November the first phase of the FTT project was finalized with a multidisciplinary meeting. A new checklist, with the five points, was introduced, which was implemented in December 2017.
The project was formally closed with a presentation to the medical staff of the ADRZ during a multidisciplinary gathering about process management. Continuation and expansion of the project are expected. There are many more patient groups for which reduction of LOS is desirable, i.e., patients with pneumonia and patients that had a Cardio Vascular Accident (CVA) with brain damage.
When the calculations of Berden are followed, an expected reduction of 250 hospital days, leads to a reduction of 2.0 nursing staff and one bed can be permanently closed. The cost reduction is around € 100.000 when is effectuated (Berden, Berrevoets et al. 2016). The other way around does not work: only reducing staff and beds does not lead to more efficiency. In a hospital cost efficiency follows quality improvement (Heuvel, Bogers et al. 2006). Medical processes are resilient to change. The only way to improve quality is doing an intervention by using an improvement model, such as LSS, combined with well thought out Change Management (ten Have, ten Have et al. 2013).
Following this project, the next step should be close cooperation with the receiving rehabilitation facilities. It is imperative that the waiting time for actual transfer is reduced from four to two days. This requires seamless medical data exchange, for the
rehabilitation physicians to be entirely, and timely, informed.
In a separate follow-up project, the five-point checklist could be made generically. The checklist is then used the second day for all patients that were acutely admitted. The second day mostly a diagnosis is formulated, and a treatment proposal is made in case of neurological, internal and lung patients. The surgical patients will mostly have had their operation, and the second day rehabilitation is formulated. The follow-up project will be carefully drafted with the principles of Change Management in mind, to increase the chances of success.
J. Results
The population of this study consisted of 97 TGHF patients that underwent surgery from July 1st to October 31rd 2017. These patients are over 70 years of age, have a hip fracture, and an explaining trauma (fall). The four-month period was divided into two cohorts of two months.
Of all patients, LOS was known. TN’s were not involved with all patients, not all patients needed a notification for rehabilitation at a center, sometimes they would go home or to a care-hotel. As both the home-care organizations and the care-hotel could provide adequate care within short notice, it is hardly discernible from available hospital data when the home-care organizations or care-hotels were involved, because the family or GP would arrange for this. The date of notification of a rehabilitation center was
documented by the TN’s.
The mean reduction in time to the announcement for rehabilitation is 1.2 day shorter in cohort 2, which is significant. A reduction occurred from 4.9 to 3.7 mean days after admission; P = 0.027 in the ANOVA test (Table 12). Because the time of announcement is normally dispersed, ANOVA is applicable.
The mean reduction in LOS of 1.1 days is not significant. The median reduction in LOS was from 8.9 to 7.8 days; P = 0.948 in the Kruskal Wallis test (LOS is log-normal
dispersed, which requires Kruskal Wallis test).
Table 12: An overview of two cohorts of all TGHF patients (N=97) that underwent surgery from July 1st to October 31. Not in every case a TN was involved, not in every case a rehabilitation center was notified.
Cohort Announcement Rehabilitation
mean days LOS, all patients mean days July - August 2017 4.9 (N= 36) 8.9 (N= 45) September - October 2017 3.7 (N= 33) Significant ANOVA P= 0.027 7.8 (N= 52)
Age profile,
The age of the patients in the two cohorts does not differ. The two groups are very comparable according to this parameter. Both cohorts had a mean age of 85 and were normally dispersed. There is a cutoff point: all patients were at least 70 years of age, because of the definition of ‘geriatric’.
Mortality rates,
In the first cohort from July-August 2017, 2 patients died within 30 days (4%). In the second cohort from September-October 2017, 3 patients died within 30 days (6%). The numbers are too small to allow for any statistics, but they are both around the 5% mark. It is concluded that 30-day mortality in these two cohorts is about the same. (see Appendix C: mortality)
GTU rate,
In the second cohort, there were somewhat more patients nursed at the GTU directly postoperatively (22 of 52 = 42%) as in the first cohort (15 of 45 = 33%). The difference was not significant (Chi-square P= 0,363) (see Appendix D: Use of GTU).
Statistics for patients with LOS > 6.0
51 patients had a LOS over 6.0. In this subcategory, 3 patients had complications: their mean LOS was 8,2. The complications all occurred to be infections. One patient went home, one patient went for rehabilitation, and one patient went back to the NH.
This extended stay for complications induced about seven extra hospital days (3 * 2.2 = 6.6 extra days). This causes only 2% of the accumulated extra hospital days after a LOS longer than 6.0 days.
When sorting all 51 patients after their destination following discharge, it occurs that waiting for a psychogeriatric place in an NH causes the longest individual delay. Waiting for a rehabilitation place induces most extra hospital days. Details are given in Table 13. The delays were nearly always due to waiting for transfer to NH’s and rehabilitation facilities. Noteworthy is that if patients that already live at an NH are longer in the hospital than three days, their place is gone and they have to wait for a new place at their old NH.
Due to rounding off totals are different than the added separate numbers in the columns. Because the difference in LOS between the cohorts was not significant, no subdivision is made in this comprehensive overview.
Table 13: overview TGHF patients (N=51) with a LOS of > 6.0 days, according to their destination after discharge, sorted after their contribution to extra hospital days.
Destination Number of
patients Percentage all patients (N=97)
Mean LOS Extra
hospital days extra hospital daysPercentage of
Rehabilitation facility 39 40% 11.9 229 70% New NH place 4 4% 25.4 78 24% Return to NH 2 2% 12.4 13 4% Home (with home care) 4 4% 7.5 6 2% Care Hotel 2 2% 7.2 3 1% Total patients 51 53% 12.3 328 100%
V. Critical reflexion, conclusions and recommendations
Introducing a Care Pathway gives rise to better quality, but does not always lead to
measurable cost reductions (Centen, Wijnen et al. 2011). At the ADRZ also some benefits evolved from the implementation of the TGHF Care Pathway, but it did not change LOS, measured after 1 1/2 years.
When studying the TGHF peri-surgical process, it turned out that patients with an uneventful recovery could be discharged after LOS of four days. Due to delays, there was an extra waiting time of around four days. The worst category patients, the ones from NH’s, had the shortest LOS of 3.4 days. At the ADRZ it proved possible to release this specific NH-TGHF category in a short time. The hospital would like to see most patients leave the fourth day after admission.
The first part of the delay in transfer to a rehabilitation center was attributed to ‘wait and see’ the first 48 hours after surgery, before sending out a transfer notification. This period was taken to see whether complications occurred. After notification of the rehabilitation, the centers needed around four days to find an available ‘bed’.
The intervention in this project was aimed at the internal processes of speeding up the announcement of an eligible patient to a rehabilitation center. The central research
question was: can we speed up the signaling of eligible patients to rehabilitation centers? A. Earlier announcement for rehabilitation
There were two significant reasons to start with adjusting the internal processes and then involve the external organizations to speed up their acceptance trajectory in a follow-up project. The first reason was that internal processes are better influenceable than external processes. The second reason being that the ADRZ was perceived as untrustworthy by external partners.
From the external stakeholder analysis, it occurred that the medical information from the hospital sometimes ‘downscaled’ the need for rehabilitation care. This lead to the conclusion that the internal processes and medical information had to be improved. Accurate and transparent internal processes are a sound basis to start discussions with the rehabilitation centers on their improvement possibilities.
Rehabilitation centers are very interested in TGHF rehabilitation patients if they can go home after a few weeks because they are a profitable patient group. A prerequisite is that this group is adequately identified and the level of needed care is trustworthily
The project was aimed at speeding up the announcement to the rehabilitation facilities. To involve the TN at first opportunity, a ’Fast Track Transfer’ (FTT) was designed. A
questionnaire was filled out by the Trauma surgeons and PA’s, and the TN were informed as soon as possible.
All rehabilitation centers in the Zeeland region had, previous to the project, approved this FTT. This included a preliminary announcement for rehabilitation to the centers of FTT patients the first or second day postoperatively, instead of the normal third or fourth day.
The first cohort, without intervention, ran in July-August 2017. During this time the intervention was prepared as defined in DMAIC phase 6. The first of September 2017 a checklist was introduced at the first ward round postoperatively (around the second day of admission). There was a measurable, significant improvement in mean waiting time before the announcement for rehabilitation to a center (for details see Appendix A). Overall there was a significant reduction in time before the patient was announced from 4.9 days to 3.7 days (ANOVA P= 0,027). There was a reduction in mean LOS from 8.9 to 7.8, which was not significant. The project proved to speed up announcement for
rehabilitation for the whole group TGHF patients.
When looking at the causes for a LOS of over 6.0, complications are a minor contributor. There were 51 patients with a total LOS of over 6.0, in this subgroup mean LOS was 12.3 days. In this group, only three patients had a complication after surgery in the form of an infection. These three patients had a LOS of 8.2, complications do not lengthen LOS very much and induce a minor amount of seven extra hospital days (3 * 2.2 = 6.6 extra days). The most extra days are caused by the patients waiting for a rehabilitation place. 39 patients waited on average 5.9 extra days (after the first regular 6.0 days) for transfer, which induced 229 extra hospital days, which contributed 70% of extra hospital days after 6.0 days stay.
Patients that could not return home anymore, also not after rehabilitation, waited the longest for transfer. These patients were eligible for permanent residence in an NH. Four of these patients waited on average 19.4 days extra, which induced 78 extra hospital days, which contributed 24% of extra hospital days.
It is concluded that nearly all patients could be discharged after 6.0 days. The current longer LOS is mainly due to waiting for rehabilitation and permanent NH placement.
B. ‘Change Management’ for medical routines
The administrative procedure proved influenceable. A checklist for the first ward round was introduced and was provided at the ward rounds by the administrative staff. The earlier announcement of the patients to external rehabilitation centers occurred. Although doctors and PA’s were involved in the change process in every step of this project, the medical process did not change.
When the surgeons and PA’s were confronted with their inability to change their working processes, they admitted they found changing their routine very, very difficult. The two interviewed trauma surgeons and two PA’s told that they could not remember what had to be changed when performing their daily routine. They only became aware afterward that they had forgotten to tell the patient and family of imminent discharge. Changing old routines around postoperative care for TGHF patients proved resilient, although every step was described in the TGHF Care Pathway and facilitated by the dedicated TGHF EPD.
When medical professionals are empowered, they can influence and improve their work processes. Multidisciplinary meetings can help establishing optimized Care Pathways.
Change Management in organizations enables dedicated attention, to change resilient
medical processes.
For project leaders of a change process, it is important to ‘make the complex simple’. Healthcare is complex, and many different medical, nursing and administrative processes are intertwined. Professionals cannot unravel these processes by themselves. A new, simple way of working has to be designed and implemented. However, the project leader can only accommodate for the professionals to make the change themselves; the
professionals cannot be forced externally or simply by designing a new digital form. Albert describes that besides these meetings collaborative communication expectations are essential, as well as supporting environmental conditions for regular and spontaneous discussions across networks (Albert, Malloch et al. 2013).
Polanyi has described ‘tacit knowledge’, which consists of actively acquired knowledge combined with experience, which has been integrated into one’s subconscious thinking, decision making and acting pattern (Polanyi 1966). When
performing a highly skilled task, such as surgeons and nurses do, most decisions about ‘what next to do’ follow automatically from this ‘tacit knowledge’.
Ten Have has described how change processes work. One of the five key points
mentioned in the Promic 2.0 model is ‘focus’. Routines can be very deeply automated, to the point that they are ‘unconsciously’ performed. When not properly addressed, i.e. by referring to a ‘critical incident’, old routines stay the same (ten Have 2011).
When a professional wants to change a routine, he or she first has to become aware of the current process. Reminders during the administrative process can increase
awareness of the recent changes. Being aware of conventional processes tends to slow down the task. In the case of the TGHF pathway, the professionals were asked to
remember the new way of working, while they were talking to the patient and making decisions. There was no reminder about the changed procedure. The medical
professionals reported that they found it difficult to change their routine from memory. The 48 hours ‘wait and see’ was a very resistant habit. In earlier days there was a rule that patients could only be announced to a rehabilitation facility, or NH if the medical treatment had finished. Although this rule does not apply anymore, and the rehabilitation facilities had approved of an ‘early warning system’, the professionals felt unsure to defer from their old routines. If the TN’s were involved one day earlier, they made the
announcement also one day earlier, so postoperative waiting time for the announcement was shortened. The 48 hours wait and see stayed the same. An in-depth discussion with the TN’s led to the conclusion that ‘old routines’ are hard to change, although there is no current rational reason.
From this study followed that timely informing of TN leads to an early announcement for rehabilitation. It is concluded that administrative procedures can be adapted.
However, the professional side of healthcare delivery is very resilient to changes. To
achieve further improvement of the hospital work processes, particular focus is necessary to effectively change the routine. In this case, to achieve an early announcement for each TGHF patient to the rehabilitation center and further reduce LOS.
C. TGHF checklist first postoperative ward round
The first recommendation from this study is to update the checklist on the first
postoperative day (second day of admission). To achieve the ultimately desired LOS of mean five days, this checklist ensures that the TN is involved the first day postoperatively. From this study followed that early involvement of the TN speeds up administrative
processes. Also an ‘expected discharge date’ is established. It is expected that communication of an expected discharge date for all TGHF patients, speeds up the discharge process and reduces LOS.
