1
________________________________________________________________________________________________________________
Influencing patient flow across hospitals and
rehabilitation care providers: a case study
_________________________________________________________________________________________________________________Master Thesis Supply Chain Management
2
Acknowledgement
3
Abstract
4
Table of contents
1
Introduction ... 5
2
Theoretical background ... 7
2.1
Patient flow... 7
2.2.
Patient flow variability ... 8
2.3
Patient flow characteristics ... 8
2.4
Patient flow coordination ... 10
2.5
Patient flow performance ... 11
2.6
Conclusion and conceptual framework ... 13
3
Methodology ... 15
3.1
Research design ... 15
3.2
Research setting ... 15
3.3
Data collection ... 16
3.4
Data analysis ... 18
4
Findings ... 20
4.1
Actual patient flow ... 20
4.2
Variability of the patient flow ... 22
4.3
Flow characteristics in patient flow ... 25
4.4
Coordination of the patient flow ... 27
4.5
Main factors influencing the patient flow performance ... 34
5
Discussion ... 42
5.1
Variability and patient flow performance ... 42
5.2
Flow characteristics and patient flow performance ... 43
5.3
Coordination and patient flow performance ... 44
6
Conclusion ... 46
6.1
Theoretical implications ... 46
6.2
Managerial implications ... 46
6.3
Limitations and future research ... 46
7
Bibliography ... 48
8
Appendices ... 50
Appendix A: Interview protocol ... 50
Appendix B: Field notes ... 58
Appendix C: Coding tree ... 60
5
1
Introduction
Nowadays, hospitals face more and more pressure to direct patients downstream the healthcare
supply chain. The increasing aging population combined with the trend of living at home longer,
increases the complexity of care problems for patients (NIVEL, 2016). As a consequence,
downstream organizations, such as rehabilitation care providers, face problems with the inflow
of patients into their organizations. For this reason, patients in the hospital stay longer than
necessary because no beds are available in the follow-up process in rehabilitation care (AD.nl,
2017). As a consequence, hospitals face higher costs and patients have to wait for the start of
their rehabilitation. This trend creates the need to come up with ideas and solutions that can
overcome these patient flow problems across the healthcare supply chain (Gittell & Weiss,
2004; Haraden & Resar, 2004).
To date, research on improving patient flow has mainly focused on individual departments,
such as emergency departments (Miro et al., 2003;
Saghafian, Austin, & Traub, 2015) and
improving patient flows hospital-wide (Drupsteen, Van der Vaart & Van Donk, 2013). These
studies show that both efficiency and quality of care can be improved at the same time in patient
flows (Villa, Prenestini & Giusepi, 2014). Little is written about improving patient flows across
healthcare organizations (Yergens, 2015), while patient flows from one organization to the
other are seen as one of the weakest spots in the supply chain (Meijboom, Schmidt-Bakx &
Westert, 2011). This study focusses on patient flows across healthcare organizations from a
supply chain management perspective.
In their study in the field of supply chain management De Vries and Huijsman (2011) describe
that patient flow research across healthcare organizations should be focussing on (1) planning
and control decisions on matching supply and demand, (2) decisions on variability and
complexity of demand and (3) coordination across members of the healthcare supply chain.
However, because this phenomenon is new and not fully understood, it is still unanswered how
these decisions need to be taken and how patient flows across members of the healthcare supply
chain can be coordinated.
6
This study addresses the following question: ‘How do variability, flow characteristics and
coordination across hospitals and rehabilitation care providers influence patient flow
performance?’
An exploratory case study will be done to get an in-depth understanding of patient flows across
hospitals and rehabilitation care. This research contributes to theory and practice. First, this
research contributes to theory in the field of patient flow across healthcare organizations, by
indicating how variability, flow characteristics and coordination influence patient flow. Second,
research on this topic gives managers insights in how to deal with patient flows across
healthcare organizations.
7
2
Theoretical background
In the theoretical background, the existing theory about patient flow will be discussed. First,
the topic of patient flow will be introduced by looking at existing streams of literature. Next,
aspects that influence patient flow will be discussed. Subsequently, the concept of patient flow
performance is introduced and possible links between the aspects of patient flow and patient
flow performance will be further elaborated on. Lastly, a conceptual framework will be
proposed.
2.1
Patient flow
Improving patient flows is known to be important for increasing performance of healthcare
delivery processes (Villa, Barbieri & Lega, 2009). Two streams of existing literature can be
distinguished in patient flow research according to Villa et al. (2014). The first stream of
literature focusses on improving patient flow with optimization and scheduling models. These
studies often focus on quantitative problem-solving tools. The second stream of research has a
more qualitative character and focusses on theory building and identifying causes of patient
flow problems.
Another observation in patient flow literature has been made by Drupsteen et al. (2013). They
argue that despite that improving patient flows in a single organizational department could
impede system-wide patient flow (Haraden & Resar, 2004), literature still focus on single
departments in the supply chain. Drupsteen et al. (2013) and Villa et al. (2014) emphasize that
patient flow research should take a system-wide approach, taking into account the patient flow
from the start till the end.
General findings on factors influencing patient flow
8
patient flows. Therefore, factors that are agreed upon in literature by influencing patient flow
are described in the following paragraphs.
2.2.
Patient flow variability
The perspective of ‘flow’ as a unit of analysis is used to see and understand how a process
works when it crosses functional boarders (Lillrank et al., 2011). A crucial factor for patient
flow research in healthcare is to understand different types of variability because variability is
one of the most important factors in disrupting flow performance when managed badly
(Roemeling, Land & Ahaus, 2017). In practice, organizations often focus on reducing obvious
waste rather than reducing variability in the process (Roemeling et al., 2017). Being able to
distinguish between different types of variability can contribute to improve effectiveness and
efficiency in the patient flow and reduce costs while improving or not reducing the quality of
care (Litvak & Long, 2000). The first type of variability is ‘natural’ variability, which is
inherent to the healthcare process because of differences in complexities of disease and
differences in the randomness of arrival (Haraden & Resar, 2004). Litvak & Long (2000)
indicated that natural variability can be divided into three types: clinical variability (variation
in the type and complexity of diseases), flow variability (the randomness in arrival) and
professional variability (variation in abilities of practitioners and healthcare delivery systems).
The second type of variability is ‘artificial’ variability, which is caused by variation in
preferences and beliefs of clinicians (Haraden & Resar, 2004) and dysfunctional management
(Litvak & Long, 2000). The key to improving flow is to remove ‘artificial’ variability and
optimally manage ‘natural’ variability in the process (Haraden & Resar, 2004).
Schmenner and Swink (1998) made an alternative distinction in improving patient flow. They
distinguish three barriers that must be dealt with to improve patient flow performance:
non-value-added activities, bottlenecks and variability. Removing these barriers can increase the
speed and reduce variance and thus enhance patient flow performance (Fredendall, Graig,
Fowler & Damali, 2009).
2.3
Patient flow characteristics
2.3.1
Flexibility in patient flows
9
care. In healthcare processes, acute care can be characterized as unpredictable, which makes
planning and scheduling of resources more difficult to nearly impossible. As a result, the
decision-making flexibility for resource allocation is diminished. The example of Aronsson et
al. (2011: 179) indicates why flexibility is needed in the patient flow: ‘The time for performing
a task during the operation may vary depending on unforeseen complications. These variations
are genuinely unpredictable and must be handled by flexibility’. In an elective setting where
resources can be scheduled and the decision-making flexibility is higher, patient flows can be
designed to maximize output and efficiency (Vissers & Beech, 2005). The differences in
predictability of the care needed at a given time can vary across healthcare processes. Therefore,
understanding what and how decisions about resource flexibility should be taken in the acute
and elective care setting can help to respond to the variability in the healthcare process.
2.3.2
Pooling in patient flows
Another way for a healthcare organization to adapt to the variability in the patient flow is
pooling. Pooling is defined by Vanberkel, Boucherie & Hans (2012: 372) as: ‘Pooling of
customer demands, along with resources to pool to fulfill those demands’. Pooling has often
been used to make patient flows more patient-centered by pooling beds, OR’s, staff and
equipment in different departments (Villa et al., 2009). In a study about the influence of pooling
on waiting times from Vanberkel et al. (2012) is shown that waiting times of centralized pooling
systems can be shorter than decentralized unpooled systems if a limited amount of services has
to be delivered. However, pooling capacities is not always beneficial, as different services that
have different performance targets could better be exposed to dedicated resources (Vanberkel
et al., 2012). The trade-off that needs to be made by organizations is that the pool is large enough
to fulfill all demand while the tasks are narrow enough to be executed (Villa et al., 2009).
2.3.3
Bottlenecks in patient flows
10
when there are fluctuations in peaks of demand, which can only be reduced by eliminating
variability. As the healthcare sector is known for the high amount of variability in patient flows,
often more than one bottlenecks could arise in the system as a whole (Vissers & Beech, 2005).
As mentioned before, elective and acute patient flows vary in the extent of variability and
predictability, which can lead to bottlenecks in the overall system. Vissers and Beech (2005)
indicate that making a distinction in these patient flows, by handling these patient flows
separate, can overcome those bottlenecks.
Identifying the bottlenecks in the patient flow is important to understand what capacities are
required for the patient flow and can be disruptive for the patient flow performance when they
are not managed. Therefore, by studying what bottlenecks are present in the patient flow can
influence how the patient flow should be managed and how the bottlenecks can be overcome.
2.4
Patient flow coordination
Coordination is another agreed upon factor that influences patient flow. Coordination is defined
by Malone & Crowston (1994: 87) as ‘the process of managing dependencies among activities’.
Coordination across healthcare organizations of the patient flow could lead to a better
performance across the healthcare supply chain (De Vries & Huijsman, 2011). However,
healthcare organizations still face difficulties in achieving a good patient flow coordination. A
lack of coordination is seen by Villa et al. (2014) and Vissers, Bertrand & De Vries (2001) as
a major problem in patient flow, as it is difficult to balance capacities and resources across
different specialties. This complexity in balancing has a negative effect on patient flow as it can
lead to bottlenecks and resource-utilization problems.
Chan & Chan (2005) indicate that central patient flow coordinating facilities could be a solution
to achieve better coordination across a network of organizations. This role in the coordination
of patient flow across the supply chain is known in the logistics and industrial setting as
third-party logistics (Selviaridis & Spring, 2007). In this case, a logistics service provider acts as a
point of contact within the supply chain and functions as a coordinator. The use of an additional
party that functions as a central coordinator might also help healthcare organizations in the
patient flow coordination.
11
of information handover and withholding of information) and (3) ineffective information
technologies (ability to support internal and external flow activities at the same time).
Ineffectiveness in coordinating activities disrupts the patient flow. Examining how these
ineffectivenesses influence the patient flow and how the inefficiencies can be reduced to
improve the coordination of the patient flow can be beneficial in this research.
2.5
Patient flow performance
Now that there is an indication of four important characteristics that could influence patient
flow performance, it is necessary to look at what is already known about patient flow
performance and how to measure patient flow performance. Existing literature is still indecisive
about the exact definition of patient flow performance and how to measure the performance of
the patient flow, while not many studies are conducted that take patient flow performance into
account.
A very broad definition of patient flow performance is formulated by Drupsteen et al. (2013).
This study indicates that patient flow performance can be measured by the time it costs to
transfer a patient from one step in the care process to the other. The study of Devaraj et al.
(2013) tries to give a deeper understanding in how to measure patient flow performance by
consistency of throughput times, timeliness in delivery and safety of the patient. This implies
that making the patient flow more steady, rapid and save could increasing the patient flow
performance.
A study of Villa et al. (2009) focused on measuring performance of patient flows logistics and
found that patient flows that are not controlled, have a negative effect on the performance of
the healthcare organization. Besides that, also the quality of care could be influenced negatively
by a poor patient flow, as inefficient and ineffective patient flow could expose patients to risks,
infections and other medical complications (Devaraj et al., 2013). Poor patient flow
performance could even lead to a higher patient mortality and a higher mortality because of
complications (Villa et al., 2009).
12
2.5.1
Operationalization patient flow performance
Existing literature does not show operationalized measures that can be used to measure patient
flow performance across healthcare organizations. However, in emergency departments
literature already elaborated on certain key performance indicators (KPI’s) that are part of
measuring the performance. In figure 1, a list of KPI’s that measure performance in the
emergency department is shown (Derived from Wilson & Nguyen, 2004).
Figure 1: Key performance indicators from the emergency departments
13
Table 1: Patient flow performance indicators
Patient flow performance indicator (derived from Devaraj et al., 2013)
Key performance indicator ED (derived from Wilson & Nguyen, 2004)
Operationalized patient flow performance indicator used for this research
Consistency Time from disposition
decision to departure
Time from decision of discharge to admission of the rehabilitation care provider
Timeliness Hours of diversion Hours of diversion of agreed and actual discharge and admission
Safety Percentage incomplete
treatment
Percentage of patients where complications occurred during the transfer of the patient
This table shows the operationalized variables that can measure the patient flow performance
across hospitals and rehabilitation care providers, derived from key performance indicators
from the emergency departments. The consistency of the patient flow can be measured by the
time from the decisions of discharge of the patient to the admission of the patient to the
rehabilitation care provider. An overall reduction in this time can have a positive effect on the
patient flow performance. Next, the timeliness can be measured by the hours of diversion of
agreed and actual discharge and admission. The timelier the patient is discharged and admitted,
the better the patient flow performs. Lastly, the safety of the patient flow is measured by the
percentage of patients that have complications during the transfer from the hospital to the
rehabilitation care provider. Fewer complications in the patient flow indicate a safer patient
flow, which increases the patient flow performance. Altogether, Bhattacharjee & Ray (2014)
has indicated that a continuous trade-off between the consistency, timeliness and safety needs
to be made by balancing the effectiveness and efficiency of the patient flow and the quality of
care provided. Therefore, improving the consistency and timeliness can lead to a decrease in
the safety and therefore have a negative effect on the overall patient flow performance.
2.6
Conclusion and conceptual framework
14
coordination across the patient flow is related to the interaction, information handovers and
information technologies, which could have on the performance of patient flows (Abraham &
Reddy, 2009). Lastly, the concept of patient flow performance that can be distinguished in the
consistency, timeliness and safety of the patient flow.
All in all, the literature on how variability, flow characteristics and coordination across hospitals
and rehabilitation care providers influence patient flow performance is still not fully
understood. Therefore, this research will focus on investigating relationships in these concepts
and what their influence is on patient flow performance. The concepts are shown in the
conceptual framework (see figure 2).
Figure 2: Conceptual framework
This conceptual framework shows the aspects of the concepts that are important to investigate.
Literature shows that these elements have an influence on the patient flow across hospitals and
rehabilitation care providers. The theoretical background is important to answer the research
question posed in the introduction: ‘How do variability, flow characteristics and coordination
across hospitals and rehabilitation care providers influence patient flow performance?’. This
15
3
Methodology
This research study aims to achieve a better understanding of patient flows across hospitals and
rehabilitation care providers. In this section the type of study that is performed will be
explained, how the data was collected and how it was analyzed.
3.1
Research design
The main purpose of this study is to identify which factors influence patient flow across
hospitals and rehabilitation care providers. Little knowledge on this topic is received and
therefore this research will opt for an inductive case study approach as called for by Karlsson
(2016), as the variables of the phenomenon are still unknown. A case study is very useful as it
allows to ask why, what and how questions with a full understanding of the phenomenon
(Karlsson, 2016; Yin, 2009). It is important to perform a case study for this phenomenon, as
multiple factors that are interrelated can be studied more in-depth to get rich and detailed
evidence in what factors influence patient flow across hospitals and rehabilitation care
providers. Besides that, this inductive case study approach could expose new factors that are
important in the understanding of the phenomenon. Furthermore, doing a case study gives the
opportunity to use multiple sources of data, such as interviews, field notes and documents that
are relevant for this study
(
Voss, Tsikriktsis & Frohlich, 2002).
To study this phenomenon the unit of analysis is a patient flow across a hospital and a
rehabilitation care provider. A single case study was performed across a hospital and
rehabilitation care provider in the northern part of the Netherlands, together with a coordinating
party. This single case can be characterized as a typical case, as the hospital and rehabilitation
care provider are one of some healthcare organizations that are part of the coordinating network.
According to the whole of the Netherlands, this case could be seen as a unique case, as the
coordination for follow-up care is provided by an external party. Therefore, the coordinating
activities are not performed by the hospital and rehabilitation care provider, but this is the
responsibility of the coordinating party. By having a separate, independent party that is solely
focused on the mediation of the patient, it could be expected that this party has a positive
influence on the patient flow performance.
3.2
Research setting
16
in trauma, elective care, amputation and acute revalidation care. The rehabilitation department
treats around 360 patients per year, varying from minor injuries to highly complex
rehabilitation. Additional information about the research setting can be found in table 2.
Table 2: Characteristics of research settingHospital Coordinating party Rehabilitation care provider Function Careful answering of
the care request of the patient
Patient mediation from the hospital to the rehabilitation care provider
Rehabilitation with the goal of returning home
Employees ± 3000 < 10 ±1500
Interviewees 3 1 5
An additional party in this single case research is the third-party logistics provider. They have
a coordinating role across healthcare organizations as they provide capacity insights in the
different healthcare providers. Their goal is to increase the overall quality of care for patients
and improve the logistic processes across healthcare organizations, with a reduction in the
overall transfer time from the hospital to the rehabilitation care provider. This coordinating
party is active since the start of 2015, which is now three and a half years. In this time, multiple
organizations have joined the cooperative. More than 25 healthcare providers (hospitals as well
as rehabilitation, home care and other long-term care providers) are connected to this institution.
3.3
Data collection
For this single case study, multiple data collection methods were used. First, field notes were
made by visiting the parties involved in this single case to get an overview of how the patient
flows across the healthcare organizations run in practice. Taking field notes created a more
in-depth understanding of the actual process, the way of working and the type of employees that
deliver the services in the patient flow. Next, to taking field notes, the visits gave information
with whom it was important to take interviews. Based on the information, Table 1 shows the
interviewees with a description of the function and the word count of the interviews.
Table 2: Interviewee characteristics
# Organization Function Word count transcript
1 Hospital Transfer nurse 5663
2 Hospital Teamleader transfer department 4846
3 Hospital Capacity manager 5145
4 Coordinator Teamleader coordinating party 6337
17
6 Rehabilitation Rehabilitation specialist 4161
7 Rehabilitation Customer service transfer nurse 5293
8 Rehabilitation Teamleader rehabilitation department 5244
9 Rehabilitation Manager rehabilitation department 5082
After the selection of the interviewees, 9 semi-structured interviews were conducted with
multiple interviewees of the different organizations. For these interviews, an interview protocol
was used to increase the validity and reliability of this research (see Appendix A). At the start
of the interview, the interviewee was asked to fill out the permission form to emphasize the
confidentiality of the interview and ask permission to record. The recordings were used to
transcribe the interviews and increase the reliability of this study. The transcripts were returned
to the interviewee to check for validity of the transcribed data and to reduce the observer bias
(Yin, 2009). Next, the interview protocol consists of multiple questions that were formulated
based on the subjects described in the conceptual model. Table 3 shows the subject, examples
of main questions and examples of probing questions that were asked during the interviews.
Table 3: Summary interview protocolPart Main subject Example main question Example sub- question A Introduction What is your relationship with regard to this
research?
What is your function with regards to the patient flow?
B Actual patient flow
I show you my findings on the patient flow so far. Is this patient flow complete and correct?
What is your task in this patient flow? What value do you add for the patient?
C Coordination How is the collaboration with department X? How clear are the mutual responsibilities?
D Variability How do you manage the complexity in care needs in the patient flow?
What is the influence of this type of variability on how the patient flow performs?
E Flow
characteristics
To what extent is it possible to be flexible in the tuning of supply and demand?
How can the flexibility make the flow more predictable?
F Patient flow performance
Which factors are most important in the patient flow?
What do you think about consistency, timeliness and safety as indicators of the patient flow performance?
18
Finally, quantitative data about the patients in the elective and acute care paths was obtained
for further analysis. Information about length-of-stay, type of disease, DRC-code, date of
admission, date of discharge and other information could be used to distinguish distinct groups
that have similar characteristics. This information is confidential and will be used for the
analysis. However, to protect the confidentiality of the patients, this information is not used in
further analysis.
3.4
Data analysis
The data that was collected from the written field notes and recorded semi-structured interviews
have been transcribed and uploaded into Atlas.ti, a coding program. An inductive way of coding
was used, to get a deeper insight into the phenomenon, as rich data was collected (Karlsson,
2016).
19
Figure 3: Coding steps analysis
Furthermore, the qualitative documents were analyzed to identify the patient flow and get
insights into how the patient flows through the process. The analysis of the qualitative
documents resulted in a flow diagram with the different steps involved in the process, to show
how the different organizations add value to the patient flow and how decisions in this patient
flow are made by the different actors in the patient flow.
To create transparency and increase the quality of this research, the validity and reliability are
addressed in this paragraph. Aspects of quality in case study research are construct validity,
internal validity, external validity and reliability (Voss et al., 2002). The construct validity is
strengthened with the use of multiple data sources (field notes, interviews and internal
documents) to see if similar results were found using these different data sources. The internal
validity is ensured using a systematic review of the achieved data by doing a coding analysis
and matching patterns in multiple stages of the coding process. Furthermore, the transcripts of
the interviews were checked by the interviewee to ensure internal validity. The external
validity, which relates to the generalizability beyond this case study is strengthened by the use
of existing literature and the use of a conceptual framework to operationalize the existing
theory on this topic. Lastly, the reliability of this research is strengthened by the use of an
interview protocol, the use of a case study database in Atlas.ti to code results from multiple
sources and the use of recordings and transcriptions of the interviews.
Table 5: Case study validity and reliability
Validity and reliability criteria Case study tactic
Construct validity • Use of multiple sources of evidence
Internal validity • Systematic coding analysis and pattern matching
• Transcript review of interviewees
External validity • Use of a conceptual model to formulate the interview protocol Reliability • Use of case study protocol
20
4
Findings
In this chapter, the results of the collected data that is found will be shown. First, in 4.1, the
actual patient flow is shown in a flow diagram. Next in 4.2 until 4.4, the collected variables
from the coding are explained and analyzed. From these variables, four main factors are taken
to be discussed, which are expressed in 4.5 in causal models. The findings will be displayed on
the basis of the concepts from the conceptual framework, coordination, variability, flow and
patient flow performance.
4.1
Actual patient flow
Based on confidential internal documents and answers from the interviews, two patient flow
diagrams are made to illustrate how the patient flows through the hospital to the rehabilitation
care provider and what the different parties across the chain do in the patient flow.
4.1.1
Patient flow elective care
21
Patient flow elective careInf or m at io n m ee tin g Pr e-op er at iv e Po st -o pe ra tiv e D is ch ar ge Tr an sf er t o re ha bi lit at io n ca re pr ov id er Patient at information meeting for elective surgery
Patient ready for discharge?
Patient discharged to rehabilitation care provider
Yes
Transfer nurse identifies opportunities for
aftercare
Transfer nurse fills in request for rehabilitation Patient operation Admission patient in hospital Patient at the nursing ward Transfer nurse controls if everything is arranged for discharge Transfer nurse sends pre-announcement Coordinating party coordinates request to rehabilitation care provider Request is send to rehabilitation care provider Rehabilitation specialist assesses patient Coordinating party sends accepted request to transfer nurse
Questions about the request Questions about the request
Beds available? Yes
No Coordinating party sends request to other rehabilitation care provider No Accept patient? Yes Yes Patient needs rehabilitation? No End: process for
rehabilitation
No
Figure 4: Patient flow elective care
4.1.2 Patient flow acute care
22
Patient flow acute care
Pa tie nt a t E D Pr e-op er at iv e Nur sin g w ar d N eur ol og y D is ch ar ge Tr an sf er t o re ha bi lit at io n ca re pr ov id er
Patient ready for discharge? Patient discharged to rehabilitation care provider Yes Transfer nurse identifies opportunities for aftercare
Transfer nurse fills in request for rehabilitation Patient transferred
to the Stroke Unit
Patient at the nursing ward Transfer nurse controls if everything is arranged for discharge Transfer nurse sends request Coordinating party coordinates request to rehabilitation care provider Request is send to rehabilitation care provider Rehabilitation specialist assesses patient Coordinating party sends accepted request to transfer nurse Questions about the request Questions about the request
Patient arrives at the ED
Yes Patient needs rehabilitation? End: process for
rehabilitation No
Beds available? Yes
No Coordinating party sends request to other rehabilitation care provider Accept patient? No Yes No
Figure 5: Patient flow acute care
4.2
Variability of the patient flow
Variability is inherent to the service delivery across healthcare organizations, which makes it
complex to optimally manage clinical, flow and professional variability and to remove artificial
variability across organizations. In this chapter, it will become clear how the organizations deal
with variability in this patient flow.
4.2.1
Natural variability: Clinical variability
23
Figure 6: Relationships in clinical variability
The increasing complexity in care needs has its influence on how the patient flow is organized
across these organizations. The transfer nurse indicates that more complex patients, for example
with cognitive or psychic problems, are harder to place, which means that it is more important
to make their handover complete. Besides that, the coordinating party is also experiencing
problems in mediation when it comes to CVA-patients, for which there is often no rehabilitation
provider in the short term. For the acute patient, it is harder to find a spot to rehabilitate, because
of the high complexity and the limited beds available. In addition, the triage (assessment of the
care needs) of an acute patient is more difficult to estimate than an elective patient after surgery.
Complex patients must therefore often stay in the hospital for 3 to 4 weeks because there is no
place at the rehabilitation care providers. When it comes to elective care ‘patients will only have
a delay in a very special secondary diagnosis, but that happens very sporadically’ (3:38). The
hospital sends a pre-announcement for rehabilitation for elective patients when the surgery is
planned in the hospital, but the rehabilitation care provider does not make use of this.
Furthermore, the rehabilitation care provider waits for accepting requests for rehabilitation after
surgery because of the variation in care needs, to see whether the condition of the patient has
changed: ‘How does the patient leave the OR? Do they leave with a delirium? If that is the case,
the picture would be very different’ (8:41). The rehabilitation care provider is not planning the
patients in front because of the increasing clinical variability and the risk of deterioration in
health during the process.
4.2.2
Natural variability: Flow variability
24
an outflow quota to the rehabilitation care provider and therefore create a more stable and
consistent flow: ‘You could add a quota, but not on this outflow. It would result in a more
consistent patient flow because there are fewer peaks’ (7:19).
The variability of the flow is seen by the hospital and coordinating party by the high number of
requests at once and by differences in time that the request is sent before the patient is ready for
discharge: ‘sometimes we get the requests on the day that the patient can already be discharged
from the hospital, this effect is disruptive for the flow, especially in complex cases’ (4:60) and
the structural differences in request arrival during the year, week and day. Generally, more
requests are sent in the fall and winter during the year, before and after the weekend during the
week and in the afternoon during the day. This variability has an effect on the coordinating task
in the patient flow, as the coordinating party often receives the requests for rehabilitation in a
high amount at once. This causes a delay in the mediation of a request to the day after and
creates a delay in the patient flow. The relationships are shown in figure 7.
Figure 7: Relationships in flow variability
Planning patients in front could reduce the flow variability and make the patient flow more
predictable, according to the interviewees. However, the rehabilitation care provider is not
planning patients in front as the health status could change after surgery, which can lead to risks
downstream the supply chain and more flow variability as the organizations have to wait before
the request for rehabilitation is answered. Therefore, the lack of planning by all the
organizations enhances the flow variability in the other areas of the patient flow. As the patient
planning in this flow is essential to make the patient flow more predictable and can have a
potential positive effect on the overall patient flow performance, the patient planning is seen as
a main factor for this research.
25
Professional variability is the variation in abilities of practitioners and healthcare delivery
systems. The interviewees indicated that there are differences in the knowledge of the execution
of the aftercare process. The transfer nurse indicates a lack of knowledge of the aftercare at the
nursing ward and by medical specialists. As the nurses and the medical specialist do not have a
clear view of the aftercare process, the way in which the information in the handover is written
about the follow-up process differs and therefore has a negative effect on the completeness of
the handover.
4.2.4
Artificial variability
The last type of variability is artificial variability, which is caused by variation in preferences
and beliefs of clinicians. The interviewees have indicated that eliminating waste is certainly
important if you want to improve the patient flow, but does not necessarily have an effect on
the performance of the patient flow: ‘of course you try to reduce having waste, but I think the
elimination of artificial variability is not specifically related to this research. You try to reduce
artificial variability at all times’ (7:22).
4.3
Flow characteristics in patient flow
In this paragraph the flexibility of the organizations, the possibility of pooling and the existence
of bottlenecks play a role in the performance of the patient flow. Organizations can be flexible
in the adjustment of capacities. Besides that, organizations can pool resources and capacities to
adapt to dependencies in the patient flow. Lastly, the existence of bottlenecks could influence
how the patient flow is designed and how this influence and limits the performance of the
patient flow.
4.3.1
Flexibility
26
Figure 8: Relationships in flexibility
First, bed flexibility is important for the rehabilitation care provider to be able to admit the
patient. The rehabilitation care provider has a system whereby beds from other departments can
be used to cope with peaks in demand. A complicating factor in being flexible with beds is that
the number of care trajectories agreed with the insurer must be taken into account of the
purchase, ‘therefore, it is important to think outside the financial boxes’ (6:51). The
rehabilitation specialist tells that the demand is generally higher than the supply, ‘the flexibility
is okay if you compare it to other rehabilitation care providers, the demand for beds is high’.
Another element that is closely related to the bed flexibility is the flexibility in the amount of
patient that can be admitted into the organization, the intake flexibility. The rehabilitation care
provider has limited capacity for the intake of patients while maintaining quality and safety
standards in the admission. The intake capacity is set at two a day but the intake flexibility is
not only dependent on the intake capacity of the organization, but also on the timeliness of the
transfer during the day (a preference for no patients in the late afternoon and on Friday), which
indicates a low flexibility in the intake.
The specialist flexibility is also essential to match supply and demand for hospital and
rehabilitation care. The rehabilitation specialist is only available to assess the handover request
once a day, instead of multiple times a day. Furthermore, the transfer flexibility is often not able
to meet all demand, which is an essential part to create complete handovers for the rehabilitation
care provider ‘As transfer nurse, you can hardly be flexible, especially on Tuesdays when there
are multidisciplinary consults. We try to spread the capacity but that is difficult, as you cannot
foresee how many requests will come tomorrow’ (3:27).
27
patient flow. This can be a result of strategic decisions, different priorities or limited resources
that the organizations face.
4.3.2
Pooling
In this case, no pooling was used across the organizations. Furthermore, the interviewees were
almost all confident that pooling capacities were not directly a very interesting solution to use
in the patient flow, because it should be costly. ‘I could not imagine what we could solve with
pooling. I think most of our capacities can be organized within 24 hours.’ (3:59). The hospital
indicated to pool personnel internally and saw pooling as a potential solution to create extra
capacity in the chain where it is needed.
4.3.3
Bottlenecks
Two broad categories of bottlenecks were found during the interviews that influence the patient
flow. First, the healthcare system functions as a bottleneck. It is said that there is a structural
overflow of patients into the system, which cannot always be handled by the hospital ‘at this
point, there are coming 40 people into the hospital while there is only space for 20 people, so
that is not going well’ (4:39). Another bottleneck is the limited resources and personnel
available. All parties agree that the performance of the patient flow is very closely related to
the available resources and staff at the right time in the right place ‘If you have a small team
like the transfer department and someone is ill, it is of course more difficult to overcome than
if you have a large nursing team.’ (7:23). The lack of resources and personnel has a direct
influence on the mediation time of the coordinating party, they indicate: ‘due to external factors,
too few beds and specialists in nursing homes, you can see that the mediation time increases.’
(2:14).
4.4
Coordination of the patient flow
The role of coordination, the completeness of the handover and insight in information in the
patient flow are main factors that can be derived from the results. In this chapter factors about
coordination, divided in interaction, handovers and information systems will be distinguished.
4.4.1 Interactions
28
care provider will take the patient: ‘Everything that the specialist in the hospital discusses with
the patient, could be a potential danger for the mediation of the patient and the creation of
wrong expectations’ (2:89). This type of ineffective interaction is caused by lack of knowledge
in the follow-up process of the healthcare professionals, as they have not specific knowledge
about the aftercare. On the other hand, the interviewees indicate that patients should be involved
as early as possible in the rehabilitation process, to align expectations and to get a complete
picture of the patient.
Figure 9: Relationships in patient interaction
The second type of interaction in the patient flow is the intramural hospital interaction, between
the transfer-nurse, the nurses of the nursing ward, medical specialists and employees of the
supporting disciplines inside the hospital. Interesting to see is the dependency of the transfer
nurse on other disciplines to gain information of the patient and wait for requests to be sent: ‘If
the nurse decides to fill in the request for rehabilitation in the afternoon instead of in the
morning, this can lead to congestion’ (4:45). The dependency on other disciplines could
influence the completeness of the handover if the other disciplines do not take their
responsibilities to make their information for the handover complete. On the other hand, the
responsibility of the transfer nurse is to achieve all the necessary information of the patient
together with communicating the aftercare plan. The coordinating party and the customer
service are doubting whether all patients are seen before this request is sent: ‘The transfer nurse
needs to interact with the patient about their rehabilitation process, but that does not happen
with all the patients’ (2:20).
29
are clear. When we need information, all that is needed is given to us, plus it is added in the
Electronic Patient Folder.’ (3:5).
The fourth type of interaction is intermural interaction, which is interaction between parties
throughout the entire patient flow. The relationships of intermural interaction can be seen in
figure 10. At this moment, tensions can be seen by the coordinator: ‘You can see that there are
tensions between the hospital and rehabilitation care provider and we are in between. We have
a role in these interactions and this has a negative effect on the cooperation’ (2:19). The
rehabilitation care provider does often react to minor inaccuracies in the correctness and
completeness of the handover, directed towards the hospital. In this case, the rehabilitation care
provider would like to see a facilitating role of the coordinating party in the interaction of
mistakes to improve the quality of the handovers. The facilitating role of the coordinating party
is not clear at the moment: ‘The mediating role between the parties is not obvious in the process’
(6:3). On the other hand, all parties indicate that they would like to have more
cross-organizational interaction, by discussing complex cases more often across the chain: ‘It is not
so black and white that it is always the transfer nurse, the rehabilitation care provider or the
nursing ward, it more complex. We should get an insight into this’ (4:19).
Figure 10: Relationships in intermural interaction
30
communicated more clearly: ‘the actual message is missing if you send the e-mail via the
coordinating party. You can indicate in a better way what the problem is if this can be
communicated directly’. Direct contact could especially be beneficial with complex cases: ‘That
one-on-one contact could be beneficial, specifically when cases become more complex’.
Furthermore, the rehabilitation specialist could speed up the assessment of the rehabilitation
request when there are questions about the request or there is something incomplete: ‘From
time to time you have the feeling that you want to have direct contact with the other
organizations so that it could be arranged faster’ (1:80).
The lack of clarity about the possibility of direct interaction is a task in the role of the
coordinating party. The coordinating party influences this as it has announced that the
organizations can interact about patient details but not about capacity. Conversely, both parties
have indicated that they want direct contact but to date, this has not yet reached a contractual
agreement.
A central factor in the coordination of the patient flow is the role of the coordinating party. The
role of this coordinating party is to mediate the patient from the hospital to the rehabilitation
care provider, while the rehabilitation care provider has all the information to make a good
decision. However, the role that has been proposed to be filled is not felt by all interviewees.
Two interviewees of the rehabilitation care provider indicate that the coordinating party does
not facilitate enough between the parties, when it comes to incomplete handovers or responding
to questions about the further transfer process of a patient: ‘I see it more as a ‘patient pass on’
organization than a coordinating party, while it could be something very beautiful. They should
be more facilitating to the whole chain’ (6:32). Also, the capacity manager of the hospital
indicates that all organizations are in a way ‘dependent on how we fill in the role of the
coordinating party as a chain’ (7:34). Therefore, the role of the coordinating party is essential
for the performance of the patient flow, when the role is not filled in a proper way, this could
be disruptive for the patient flow as a whole.
31
the information is correct. You must be able to assume that the information is correct. This
should be discussed across the organizations.’ (2:91).
4.4.2
Handovers
Many factors were found that are part of the handovers in the coordination of the patient flow.
Handover completeness, handover appointments, handover questions, handover updates,
handover writing, handover assessment and the warm handover were found as important factors
that influence the patient flow. The relationships of the handovers resulted from the coding are
shown in figure 11.
Figure 11: Relationships in handovers
The second main factor that is essential for the performance of the patient flow is the
completeness of the handover. The completeness of the handover refers to the extent to which
the information that is needed for the follow-up process of the patient is filled into the handover.
Indicated by all parties, the completeness of the handovers is one of the most important factors
that influences patient flow performance: ‘The more complete the handover is, the easier it is
to mediate the patient throughout the chain and the easier it is for the rehabilitation specialist
to judge the request’ (1:90). The completeness of the handover depends on many elements
32
high amount of risk in the patient flow and is an essential element in the performance of the
patient flow.
The appointments made about the handovers are not complied to by every party in the same
way. In addition, it is not mandatory for certain disciplines to fill in certain items in the
handover, which should become an obligation. Furthermore, all interviewees indicate that the
complicating factor starts when the care path of the patient cannot be followed because of
changes in the health status.
Handover questions are asked when information in the handover is not fully clear when the
patient is assessed to be admitted to the rehabilitation care provider. The questions are often
asked about further details of the patient, for example cognition. These handover questions do
influence the performance of the patient flow, as these questions are sent via the coordinating
party to the hospital, have to be assessed by the hospital and sent back to the rehabilitation care
provider, while the patient is still waiting. In this way, the more questions are asked about the
handover, the less complete the handover is.
Another element in the handover is the handover updates, which can be seen as an element of
the completeness of the handover. The handover needs to be updated if the health status of the
patient changes and the patient has to stay in the hospital for longer: ‘If the patient is delayed,
you have to give a correct update for the rehabilitation specialist to indicate what has happened
and what actions are undertaken in the meantime’ (1:31). The results of the interviews and
internal documents show that there are no appointments on how, when and with what
information the handover needs to be updated if the patient is delayed. Real-time information
is essential in the care process, as this prevents the professionals from making mistakes (e.g. in
medication prescription).
The next element is the handover writing. The interviewees indicated that there is a high
variation in the way handovers are written. As the nurse, specialist at the hospital and
transfer-nurse fill in the handover, differences in the style of writing, length of the handover and
handover language could influence the quality of the handover. For now, no appointments are
made about the writing style, length and language used that fits the assessors at the rehabilitation
care provider: ‘Often something else is meant that is actually written. This creates ambiguity
because the message is interpreted differently than the other has thought of it.’ (7:6). The
33
should be and how this writing can be improved: ‘We have to talk to see what is useful when it
comes to writing the handover’ (4:57).
Another element in handovers it the assessment of the handover. There are some differences in
the way the handovers are assessed. The appointment in the patient flow is that the handover
request of the patient is assessed within 4 hours by the rehabilitation specialist, however the
specialist is only available once a day for the assessment. This has an enormous impact on the
patient flow, as the specialist is not able to assess the request directly, this leads to congestion
in the process: ‘It would be better if the rehabilitation specialist could assess the requests more
times a day. That it will be assessed directly after the request has arrived.’ (6:30).
The last element of handovers is the warm handover, which is named as a potential solution to
increase the completeness of handovers and increase patient flow performance. All
interviewees, except for the coordinating party, would like to have warm handovers, as patients
could already be seen earlier in the process. However, the coordinating party has said that this
type of handover cannot be implemented because this could lead to a conflict of interest between
all organizations in the chain.
4.4.3
Information systems
The insight in information across organizations is another main factor that influences the patient
flow performance. The interviewees indicated that a basic information system is used in this
patient flow where the hospital can send their requests to and the rehabilitation care provider
can adjust the availability of beds. The coordinating party is the administrator of the shared
patient transfer information system and is the only party who has information about the
available bed capacity at each organization, but this is not visible for the other organizations:
‘The system is not at our disposal, it is not for common use. We can only enter our own bed
capacity’ (7:39). This information system is only filled with basic information, about
adjustments in bed capacities. The interviews have shown that the organizations would like to
be able to have more insight in information that they can use to monitor the patients as they
flow through the system. Next to that, the organizations add that it would be beneficial if all
organizations have the same relevant information, to facilitate the patient conversation, to plan
the patient in front and to see where the patient is located in the process ‘If that would be
beneficial for the patient flow, we should certainly not fail to do so.’ (7:37). The manager of
34
therefore increase quality. The rehabilitation care provider would then be able to plan the
patients in front. The relationships between the variables are shown in figure 11.
Figure 12: Relationships in information systems
However, there are some limitations to this solution as the capabilities of the ICT-information
system are limited and the organizations are dependent on the capabilities of the ICT
professionals to create this information sharing. Furthermore, it is mentioned that the
coordinating party does not allow information sharing about capacities. All in all, the
interviewees indicate that having insight into the information of the other organization could be
beneficial for the patient flow.
4.5
Main factors influencing the patient flow performance
From the results, four main factors are derived that are of decisive importance for the patient
flow performance: the role of the coordinating party, the completeness of the handover, the
patient planning and the insight in information. The four causal models are made to illustrate
the underlying relationships and links within the concepts of variability, flow characteristics
and coordination across hospitals and rehabilitation care providers and how they influence the
patient flow performance.
4.5.1
Role of the coordinating party
35
in the coordinating task. Figure 1 shows the relationships between the variables that determine
the role of the coordinating party.
Figure 13: Causal model - role of the coordinating party
The increasing variability in the complexity of patients and the increasing arrival variability of
requests for rehabilitation makes the role of the coordinating party more complicated. The
coordinating party was introduced to coordinate patients with different complexities. However,
the increasing complexity of patients causes that the patient can no longer be mediated in the
given time, which causes longer waiting times. Besides that, the growing variability in the
arrival of requests for rehabilitation (variety in the number of requests that is sent at multiple
moments during the day) causes the coordinating party to work overtime or answer the request
at the next day. All in all, the increase in types of clinical and flow variability show that the
coordinating party is not yet able to deal with the demand in the coordination of this patient
flow.
36
Furthermore, the task to coordinate has become not only more complex by an increase in the
variability and by the organizational flexibility, as also other factors in the organization of the
coordinating task play an important role. One part of coordination is the use of information
systems to coordinate the patient flow. The coordinating party is the administrator of the
information system, whereby this party is the only organization that has insights in the available
capacities across the patient flow. In this case, the coordinating party is not willing to make this
information available for all organizations because this could lead to conflicts of interest.
Therefore, the coordinator shields this capacity information which causes a lack of insights in
information for the hospital and rehabilitation care provider. This makes it harder for the other
organizations to plan their patients up front, as there is no clear insight in what the capacities
are in the patient flow.
Thereafter, handover appointments imply how the organizations handle the handovers in
different situations. The lack of the current handover appointments complicates the
coordinating role as the coordinating party cannot control the quality of how the handover is
filled. Furthermore, the current way of direct interaction across the organizations makes the
coordinating role more complicated. Because of the growing complexity of patients, more and
faster interaction across the hospital and rehabilitation care provider are needed to facilitate a
well-organized transfer of the patient. As the direct interaction between care providers is not
allowed or facilitated by the coordinating party, the coordinating party is not able to meet the
facilities that can handle the needed direct interaction across the organizations. All in all, the
previously mentioned shows that the coordinating party is not able to handle the task for
coordination as it could be. In addition, this increasing complexity of the coordinating role
creates a lack of trust at the hospitals and rehabilitation care providers as they are not fully
satisfied with the filling of the coordinating role.
Role of the coordinating party and patient flow performance
37
patient flow safety in the lack of handover appointments, where there is no control about the
quality of the handover and thus information can be missed which has a negative effect on the
safety of the patient.
4.5.2 Completeness of the handover
Another important factor that is related to the patient flow performance is the completeness of
the handover. The completeness of the handover is determined by and dependent on the
variability in the patient flow, the flexibility of the transfer department and a number of
handover conditions that play a role in the coordinating task. Figure 2 shows the causal model
of the handover completeness.
Figure 14: Causal model - handover completeness