• No results found

A multiple case study to explore how engagement influences the relationship between Lean maturity and second-order problem solving

N/A
N/A
Protected

Academic year: 2021

Share "A multiple case study to explore how engagement influences the relationship between Lean maturity and second-order problem solving"

Copied!
50
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The promise of Lean in healthcare:

A multiple case study to explore how engagement influences the relationship between Lean maturity and second-order problem solving

Iris Brouwer S2155176

Supply Chain Management & Technology and Operations Management

University of Groningen

Faculty of Economics and Business Supervisor: Prof. Dr. Ir. C.T.B. Ahaus

Co-assessor: Dr. G.C. Ruël

(2)

2

Abstract

There are still mistakes made in hospitals, sometimes even leading to premature death. The amount of mistakes must be reduced and that can be done by implementing a Lean programme.

This research investigated how the implementation of a Lean programme leads to an increase in second-order problem solving which in its turn influences the level of performance. In addition, the moderating role of engagement in that relationship is investigated. However, before the role of engagement could be explored we had to define engagement since there are many definitions and perspectives on it. A multiple case study and statistical analysis were used to explore the before described relationships. This research contributes by proposing four attributes that define engagement. Moreover, it shows that engagement moderates the relationship between Lean maturity and second-order problem solving. Those insights can be used by practitioners by making sure that the engagement of employees is at a high level when implementing a Lean programme. This ensures that the performance will be increased and thereby the amount of mistakes reduced.

Keywords; healthcare, Lean maturity, second-order problem solving, engagement, nurses, performance.

(3)

3

Table of contents

Abstract ... 2

Introduction ... 4

Theoretical background ... 6

Lean in healthcare ... 6

Lean maturity ... 7

Second-order problem solving ... 8

Employee engagement ... 9

Model development ... 11

Methodology ... 12

Research method ... 12

Case selection and description ... 12

Data collection ... 14

Data coding and analysis ... 15

Results ... 18

Background descriptives ... 18

Correlation analysis ... 18

Regression analysis ... 19

Within case analysis ... 20

Lean in healthcare ... 22

Cross case analysis ... 22

Discussion ... 28

Influence of Lean maturity on second-order problem solving and maturity ... 28

Defining engagement ... 29

Moderating role of engagement ... 31

Conclusion ... 33

Implications for theory and practice ... 33

Limitations ... 34

Suggestions for further research ... 34

References ... 36

Appendices ... 40

Appendix A ... 40

Appendix B ... 42

Appendix C ... 44

Appendix D ... 47

Appendix E ... 48

Appendix F ... 50

(4)

4

Introduction

In 2015, a 19-year-old patient died through mistakes made in a Dutch hospital. This led to a huge commotion since there are still patients harmed in hospitals during treatment, sometimes even leading to premature death. Although it is hard, the occurrence of these incidents could be minimized by improving the quality of healthcare and guarantee patient safety. Lean, also called Lean Thinking or Toyota Production System (TPS), might help in improving the quality of healthcare and guaranteeing patient safety. The aim of Lean is to continuously improve every process every day (Dombrowski & Mielke, 2013) and it is based on a set of technical practices that focuses on maximising value for customers (Burgess & Radnor, 2013). Lean is already implemented in a wide variety of healthcare settings, however research shows that the implementation does not always lead to an increase in performance (Mazzocato, Savage, Brommels, Aronsson, & Thor, 2010). According to Hasle, Nielsen, & Edwards (2016) the level of performance may be partially determined by the level of Lean maturity. This can be defined as the level to which Lean is adopted in the organisation or the extensiveness of Lean implementation (Canato, Ravasi, & Phillips, 2013; Wong, Ignatius, & Soh, 2014). Although literature increasingly reports the implementation of Lean in healthcare, it did not provide insight on the level of Lean maturity in healthcare. Gaining insight in this level is necessary since it might give the solution for the different levels of performance measured in healthcare organisations. Therefore, this research will focus on how Lean maturity influences the level of performance in hospitals because it can lead to a higher quality in hospitals and a safer environment for patients.

Besides the fact that Lean maturity might play a role, research suggests that the way problems are addressed also affects the success of the improvement efforts (Tucker, Edmondson, &

Spear, 2002). Literature makes a distinction between first-, and second-order problem solving.

First-order problem solving allows work to continue but does nothing to prevent a similar problem from occurring (Tucker & Edmondson, 2003). Second-order problem solving, on the other hand, investigates and seeks to change underlying causes (Tucker et al., 2002). It is more a cognitive approach which allows organisations and employees to learn by not only solving the direct problem, but also by addressing root causes and engaging in additional steps to prevent recurrence in the future (Burgess & Radnor, 2013; Tucker, 2009). This research will focus on second-order problem solving since it is expected to be in line with Lean, because it strives for continuous improvement (Dombrowski & Mielke, 2013; Tucker et al., 2002). It is important to include second-order problem solving in this research because it might play a role in the relationship between Lean maturity and performance and therefore can contribute to a higher level of quality and safety for patients in hospitals.

As mentioned before, second-order problem solving might be the bridge between Lean maturit y and performance. However, the supportive conditions for second-order problem solving are not well investigated in literature. The engagement of employees might be a supportive condition since it is well recognised in business and industry literature that the contribution of employees can play an important role in improving business and quality outputs (Harter, Schmidt, &

Hayes, 2002; MacLeod & Clarke, 2010; Xanthopoulou, Bakker, Demerouti, & Schaufeli,

(5)

5

2009). Also in healthcare, engaging clinicians from whichever health setting or discipline, is increasingly acknowledged to be an essential precondition for the success of quality improvement initiatives (Siriwardena, 2009). Especially the engagement of nurses is important since they form the largest health professional group in all nations. Besides that, they are in the best position to solve problems and eliminate root causes (Tucker & Edmondson, 2003). Also Thorp et al. (2012) showed in their research that engaging nurses is of utmost importance because they have a synergistic impact on patient safety. However, there seems to be a lack of a clear understanding and definition of engagement and more specifically nurse engageme nt.

Besides that, it is not investigated how the engagement of nurses influences the relations hip between Lean maturity and second-order problem solving.

This research aims to address the aforementioned gaps by first addressing whether Lean maturity influences second-order problem solving and performance. This brings forward the first research question. RQ1: How does Lean maturity influence the level of second-order problem solving in order to increase the level of performance? Thereafter, this relationship will be extended by the moderating role of engagement. Before investigating that, we need to explore how the engagement of nurses can be defined. This leads to the second research question. RQ2: How can the engagement of nurses be defined? The third and final question addresses the influence of engagement on the relationship between Lean maturity and second- order problem solving. RQ3. How does employee engagement influence the relationship between Lean maturity and second-order problem solving of a nurse team in order to increase the level of performance?

To answer these research questions a multiple case study with semi-structured interviews taken from nurses in two different hospitals across The Netherlands is performed. The two hospitals are at different levels of a Lean implementation programme known as the Productive Ward.

This initiative aims at increasing the proportion of time spent by nurses in direct patient care, improving experiences of patient and staff, and making structural changes to the use of ward spaces to improve efficiency (Wright & McSherry, 2013). This research focuses on increasing the time spent by nurses in direct patient care, since it can increase the quality and safety level for patients. The research questions make two different contributions to the literature. First, the definition of engagement in the healthcare setting is explored since there are many differe nt definitions and perspectives on engagement. Second, the knowledge of the moderating role of engagement will be extended, especially within a healthcare context. This could help healthcare organisations in the understanding of why it is important to engage the whole organisation when implementing a Lean programme.

This research is organized as follows. First the theoretical background will be developed based on literature about the adoption of Lean in healthcare organisations. Thereafter, Lean maturit y, second-order problem solving and employee engagement will be discussed in-depth. Then the method of data collection will be described followed by the results that are found in this research. This will be followed by a discussion of the results and the limitations will be presented. In the final section, conclusions will be drawn, implications for theory and practice are presented and finally suggestions for further research will be given.

(6)

6

Theoretical background

In this section the theoretical framework of this research will be described. First, the concept Lean in healthcare will be explained to show how successful the Lean implementations were.

Second, an explanation of the constructs Lean maturity, second-order problem solving and employee engagement will be given. Thereafter, the hypothesis and conceptual model are provided, summarizing the relationships which form the fundaments of this research.

Lean in Healthcare

Lean in healthcare has garnered considerable interest over the last decade due to the challenges regarding safety, effectiveness and value shared by most healthcare organisations (McGlynn, 2004; Spear, 2005). Although Lean is a popular concept and used in many differe nt organisations, a general agreement about its definition is missing. This lack of clarity is evident from the multiplicity of descriptions and terms used with respect to Lean (Shah & Ward, 2007).

Therefore, Shah & Ward (2007, p.791) tried to propose a definition that captures all facets of Lean and stated that: “Lean production is an integrated socio-technical system whose main objective is to eliminate waste by concurrently reducing or minimizing supplier, customer, and internal variability.” Although it seems very promising to implement Lean in healthcare, literature shows that it is questionable whether Lean implementation leads to an increase in performance (Mazzocato et al., 2010). First of all, it is hard to implement Lean in almost every organisation since it is built on tacit knowledge. This means that it is connected to the culture of the organisation, it is not written down and it is difficult to express the true meaning of it (Jimmerson, Weber, & Sobek, 2005). Spear (2005) mentioned that in healthcare, no organisation has fully institutionalised to Toyota’s level, the ability to continuously improve, systematically eliminate waste, and reduce variability and buffers. Companies that try to adopt the ‘Toyota Way’ often focus on solutions to improve processes, even though it is the ability of managers to actively facilitate learning through questions and problem solving that determines success, rather than directly instructing employees (Spear, 2005). Besides that, Dahlgaard, &

Dahlgaard-Park (2006) stated that the implementation of Lean focused too much on process improvement that it loses its perspective on people. Also Dombrowski & Mielke (2013) specified that tools and methods are essential but they cannot achieve any results in itself. The main challenge is the change in behaviour of the mindset of the employees. If this is accomplished it will result in a new way of thinking in which tools are constantly adapted to fit the ever changing needs and problems of the organisation (Ballé & Régnier, 2007). Even though the application of Lean is still growing (Burgess & Radnor, 2013), there seems to be a disproportion between the popularity of Lean in hospitals and the limited effect of it – at least so far. According to Hasle, Nielsen, & Edwards (2016) the limited effect of the implementa t io n may be partially determined by the level of Lean maturity. However, literature does not provide insight on how Lean maturity influences the performance of the implementation. Therefore, this study aims to investigate whether the level of Lean maturity influences the level of performance which might explain the limited effect of Lean in healthcare. However the question remains what Lean maturity exactly is and how it can be measured. For that reason, it will be discussed in more detail in the next section.

(7)

7 Lean maturity

As early as the mid-1990s researchers and consultants started with proposing instruments for assessing Lean maturity (Malmbrandt & Åhlström, 2013). The use of a maturity measureme nt method can be a helpful way of managing the major transformation because it provides directions, prioritizes improvement opportunities and guides cultural changes. Lean assessment models strive to measure the extent to which Lean tools and principles are adopted within an organisation (Hasle, Nielsen, & Edwards, 2016), which in its turn might explain the performance of the organisation.

When assessing the level of maturity, literature showed that there are two items that are usually measured. The first type of items focus on enablers of Lean adoption and the second focus on the extent of the use of Lean practices:

1. Enablers often include management commitment, training for employees and time and resource allocation (Shah & Ward, 2007; Soriano-Meier & Forrester, 2002). These types of items show the importance of achieving behavioural and even cultural change for the adoption of Lean (Bhasin, 2011; Boyer, 1996; Nightingale & Mize, 2002).

2. Lean practices focus on the way of working that is seen as consistent with Lean principles. Lean principles are defined by Womack & Jones (1996) and focus on specifying value for the customer, identify value stream, make the product/service flow continuously and standardise processes around best practice, introduce “pull” between all steps and manage towards perfection. Examples of the Lean practices are: the usage of process mapping (Gurumurthy & Kodali, 2009; Singh, Garg, & Sharma, 2010), the standardization of tasks (Doolen & Hacker, 2005; Nightingale & Mize, 2002; Singh et al., 2010) and the usage of visual signals (Mejabi, 2003; Sawhney & Chason, 2005;

Wan & Chen, 2008).

It is important that an instrument that assesses the level of Lean maturity includes both items (Malmbrandt & Åhlström, 2013). The enablers give a good overview of the supporting structure for Lean. It is critically important to have a supporting structure since it leads to successful adoption (Hines & Lethbridge, 2008). However, enablers do not guarantee whether there have been any real process improvement or if these have actually led to any improved performance results (Malmbrandt & Åhlström, 2013). The importance of assessing the use of Lean practices should be seen against the background of the fact that Lean adoption is a long and complex process. Therefore, it is expected that the performance will be higher when an organisation is longer active with implementing Lean. According to Malmbrandt & Åhlström (2013) performance is seen as a third important item when measuring Lean maturity. However, in this research performance is not measured as an element of maturity, but it is seen as a result of Lean implementation which can be influenced by different factors. Therefore, it is decided to not include performance, but only enablers and practices when assessing the level of Lean maturity in this research.

(8)

8

In literature, the items enablers and practices are measured with different tools and methods.

Some tools use the Likert-type scales where items range from “no adoption” to “full adoption”

(Sakakibara et al., 1993; Shah and Ward, 2007; Scoriano-Meier and Forrester, 2002). The advantage of using the Likert scale is that it permits the usage of parametric statistical methods for analysis (Hair et al., 1998). However, the main disadvantage is related to the conceptual confusion about Lean services (Shah and Ward, 2007), with the existence of a multitude of interpretations of the term “Lean” (Papadopoulos et al., 2011). Another tool for measuring Lean maturity is the Lean Enterprise Self-Assessment Tool (LESAT) developed by the Lean Aerospace Institute which uses different maturity levels for each stage (Nightingale & Mize, 2002). However, maturity levels can be difficult to define, especially if the organisation is just starting to adopt Lean. This can be solved by using the general definitions of each maturit y level developed by Nightingale and Mize (2002) and adapted by Malmbrandt & Åhlström (2013) (Table 1). Those levels

describe for each enabler and practice how well the organisation has adopted Lean.

This means that it is not necessary to imagine what the ideal state would look like in the specific organisational setting and on the specific parameter.

Therefore, this last approach will be chosen to assess the level of Lean maturity in a healthcare environment.

Second-order problem solving

Problem solving which is about identifying and resolving problems that occur in the execution of day-to-day work routines (Jones & McBride, 1990; Mukherjee, Lapré, & Van Wassenhove, 1998) plays an important role when it comes to implementing Lean. Research suggests that the way how problems are addressed affects the success of improvements. A problem is defined as an undesirable gap between an expected and observed state (Brightman, 1988; Kepner &

Tregoe, 1976) that hinders a worker to complete his or her task. Problems encountered by front- line workers, who are responsible for producing the goods or services sold by organisatio ns, can impact product quality or customer satisfaction (Tucker et al., 2002). Front-line workers, which refers in this research to nurses, are typically a large percentage of the workforce and can provide valuable leverage in both

identifying and removing problems as it is their work routines that are being disrupted by problems and they often have first-hand access to data about the causes or consequences of them (Roth, 1985). Tucker et al. (2002) explored

Table 1: Adopted from Malmbrandt & Ahlström (2013)

Table 2: Problems faced by front-line workers adapted from Tucker et al. (2002)

(9)

9

five types of problems that are faced by front-line workers during their daily work which can be found in Table 2. In sum, nurses are well situated to identify, analyse, and resolve daily problems that arise and, through this local improvement activity, to contribute to ongoing organisational learning (Tucker et al., 2002).

Literature on problem solving makes a distinction between first-order and second-order problem solving. First-order problem solving attempts to remedy the immediate problem but does not try to change underlying conditions which created the problem. Workers exhibit first- order problem solving when they do not expend any more energy on a problem after obtaining the missing input needed to complete a task (Tucker et al., 2002). This means that a quick workaround is applied to one of the five problems mentioned in Table 2, without influenc ing the probability of repetition. Second-order problem solving, on the other hand, goes further and attempts to change the system so that the process failures will not reappear (Tucker et al., 2002).

This type of problem solving is predicted to improve the systems by way of addressing root causes and making changes that prevent the recurrence of similar exceptions. However, second- order problem solving in healthcare is just a small part of all solved problems (Tucker &

Edmondson, 2003). Second-order problem solving can be displayed by three major actions;

1. The first action of second-order problem solving is communicating about exceptions to people positioned to address underlying causes. Nurses can choose to solve a situatio n on their own or to communicate to the source about the problem.

2. The second action of second-order problem solving is the effort to find and remove underlying causes of problems.

3. The last action is experimentation which involves structured trial and error episodes to test whether a potential solution produces the desired outcome (Lee, Edmondson, Thomke, & Worline, 2001).

Through the implementation of Lean, it is expected that the degree of second-order problem solving will be increased. This means that the three major actions will be observed more frequently. However, whether the maturity of Lean has an impact on the problem solving behaviour of nurses is not investigated in-depth. Therefore, this research aims to investigate how Lean maturity influences second-order problem solving which in its turn will influence the level of performance. Besides that, the moderating role of engagement will be investigated.

What engagement entails will be described in the next section.

Employee engagement

Before we can investigate how employee engagement affects the relationship between Lean maturity and second-order problem solving, we should create a clear understanding of employee engagement. It is important to create a clear understanding and definition of employee engagement since there is a range of definitions of, and perspectives on, engageme nt even within academic literature. The concept employee engagement was developed by Kahn (1990, p. 694) and he defined employee engagement as the “harnessing of organisat io n members’ selves to their work roles; in engagement, in employment of people and in expressing themselves physically, cognitively, and emotionally during role performances.” However, when the focus is shifted to the experience of the person, work engagement is defined as a

‘positive, fulfilling work-related state of mind’ (Schaufeli, Salanova, González-Romá, &

(10)

10

Bakker, 2002) and well-being at work is ‘characterized by vigour, dedication, and absorption’

(Schaufeli & Bakker, 2004). Vinje & Mittelmark (2008) defined engagement as ‘searching for, experiencing, and holding on to the meaningful work that enables one to live one’s values’

(p.200). Although all three definitions define engagement, they are quite different from each other. Therefore, this research aims to explore the meaning of engagement in a Lean environment.

The attributes are the characteristics of the concept that both define and differentiate the concept from others (Walker and Avant, 2010). Kahn (1990) described his attributes of employee engagement as: “The employment and expression of the persons’ ‘preferred self’ in task behaviours that promote connections to work and to others, personal presence and active, full role performance.” However, Schaufeli & Bakker (2004) divided engagement into the attributes vigour, dedication and absorption (Schaufeli & Bakker, 2004) and those are frequently used in work engagement research. Vigour is characterized by high levels of energy and mental resilience while working, the willingness to invest effort in one’s work, and persistence also in the face of difficulties. Dedication is characterized by a sense of significance, enthusia s m, inspiration, pride, and challenge. The third dimension of engagement is called Absorption and is characterized by being fully concentrated and happily engrossed in one’s work, whereby time passes quickly and one has difficulties with detaching oneself from work. More recently, Vinje and Mittlemark (2008) described three inter-related attributes of nurses’ work engageme nt.

Calling, which provides the path to meaningfulness, is defined by the feelings of having a mission or purpose, including commitment. Zest for work happens when experienc ing meaningfulness in work and is defined by the feelings of vocation-related joy, happiness and enthusiasm. Finally, vitality which is the ability to hold onto meaningfulness in work is defined by the feelings of vigour, strong life energy, and the will to exert oneself.

Table 3: overview of definitions and attributes

To this extent, there are several definitions and attributes used to define employee engageme nt (Table 3). As mentioned earlier, the three definitions are somewhat different from each other.

Kahn (1990) presents a more comprehensive model of engagement, but he does not propose an operationalisation of the construct. Schaufeli et al. (2002) proposed a definition as the opposite

(11)

11

of burnout in combination with the attributes vigour, dedication and absorption, which were defined by Schaufeli & Bakker (2004). While it seems that there are some similarities to the attributes of Schaufeli & Bakker (2004), Vinje & Mittlemark developed a specific definit io n for nurses and narrowed engagement down into meaningfulness.

Although a growing body of evidence supports the relationship between engagement of the employees at work and organisational outcomes, including those which are performance-based, it is never researched how engagement relates to the principles of Lean. Moreover, definit io ns and measurements of engagement at work, and more specifically nurse engagement, are poorly understood although professions in both the academic and clinical arena shows engagement as an important work-related factor. The poorly understood definition of engagement is also shown by the diversity in definitions explained in the previous paragraph. Therefore, this research will investigate how the engagement of nurses can be defined and besides that how engagement influences the relationship between Lean maturity and second-order problem solving.

Model development

The development of the theoretical background provides a basis to place the research questions into a conceptual model. In this model there are three research question situated. The first question stated: How does Lean maturity influence the level of second-order problem solving in order to increase the level of performance? This question focuses on the interaction between the different concepts with the aim to develop an understanding of the relationships. The second research question ‘How can the engagement of nurses be defined?’ focuses on exploring the meaning of engagement in this setting since there are many definitions of it. The third question

‘How does employee engagement influence the relationship between Lean maturity and second- order problem solving of a nurse team in order to increase the level of performance?’ elaborates on the first and second questions. This means that in the last question engagement acts as a moderator and second-order problem solving as a mediator. This could shed light on the role of engagement in the relationship between Lean maturity and second-order problem solving. The questions focus on the implementation of a Lean programme in hospitals with the aim to increase the level of performance. The expectation is that when a nurse team is at a higher level of Lean maturity, second-order problem solving will be more frequently observed which will lead to an increase in performance. The relationships can be found in the conceptual model which is shown in Figure 1.

Figure 1: conceptual model

(12)

12

Methodology

Research method

In this research, the case study method is used in order to find an answer to the research questions. A case study is an appropriate method for examining how and why questions with understanding of the context (Yin, 1994). This matches the questions of this research, but furthermore healthcare poses a specific context in comparison to other services (Berry &

Bendapudi, 2007). Another important element of a case study is that analysing how and why questions can lead to theory development (Voss, Tsikriktsis, & Frohlich, 2002). Since this research strives to make a contribution to theory development about Lean implementation in hospitals it shows a close fit for a case study. Overall this case study can be seen as an explorative research since the role of employee engagement in the relationship between Lean maturity and second-order problem solving has received little attention. In addition, there is no clear single set of outcomes (Eisenhardt, 1989; Yin, 2009). However, there will also be a quantitative part in this research in which the relationships between the variables are statistica l ly tested.

In order to conduct a case study, Miles & Huberman (1984) suggest to develop a conceptual framework that underlies the research. That framework explains the main focus that will be studied – the key factors, constructs, or variables – and the presumed relationships amongst them. It is important to develop such a framework because the amount of data that can be collected is vast. Therefore, the stronger the research focus, the easier it is to handle the volume of data during the case study (Eisenhardt, 1989). The research questions and conceptual model are both developed in the previous sections in order to give a clear direction. The unit of analysis is represented by the selected cases in case study research (Voss et al., 2002). In this study the unit of analysis, derived from the research questions, is a nurse team. By doing so the within case analysis can provide insights on how nurses define and understand engagement and the comparison across cases can show specific patterns between the variables. Selecting mult ip le cases to augment internal and external validity is a vital issue. Therefore, it will be discussed in the next section.

Case selection and description

In selecting the cases, it is important that the characterist ics that are identified in the conceptual model are reflected (Yin, 2009). First of all, the cases were selected based on ward teams that currently follow a Lean-based program, namely the

Productive Ward. That

programme, originally developed by the NHS, aims at increasing the proportion of time spent by nurses

Figure 2: Productive Ward house

(13)

13

in direct patient care, improving experiences of patient and staff, and making structural changes to the use of ward spaces to improve efficiency (Wright & McSherry, 2013). The programme comprises 11 modules, a toolkit and three manuals that are arranged in a structure known as the

‘The PW house’ which is showed in Figure 2. The three foundation modules are impleme nted first (Knowing How We Are Doing, Well Organised Ward and Patient Status at a Glance), followed by eight process modules that focus on fundamental aspects of nursing practice. After the foundation modules the hospital can choose which of the eight modules they want to start with. The lead time of the foundation modules vary from 2 till 22 weeks whereas the process modules has a lead time of 6-8 weeks on average. Those lead times say something about the duration of the implementation of Productive Ward. As can be seen, it is a long- term programme of which is expected that it will influence the ability of nurses to solve underlying problems. Therefore, it is possible to measure how the level of Lean maturity influences the level of second-order problem solving and how Lean maturity influences performance.

In total, we obtained 15 ward teams from two different hospitals in the Netherlands who wanted to participate in this research. Since it was not possible to select the teams by ourselves, we had to check whether the 15 ward teams showed differences in the level of duration. As mentioned before, the duration says something about how long the team is working on the implementa t io n of Productive Ward. It is important to ensure and control that we had teams in different levels of duration in order to increase the variance of the teams in this research. Table 4 is an overview of the teams classified in a duration level. Through the fact that we could not select the teams by ourselves, we have a larger group in duration level 3. However, we can see in this table that there is enough

variance in duration level which makes it possible to compare the teams with each other in the cross case

analyses. Selecting the cases based on the level of duration ensures that we will have a variety in the level of maturity since the expectation is that when you are longer working on implementing Lean you will be in a higher level of Lean maturity. By ensuring a variety in the levels it will be possible to investigate whether there is a relationship between Lean maturit y and second-order problem solving, how that relationship is influenced by employee engagement and how maturity in its turn influences performance. It can be expected that a ward team with a high level of duration is assessed as more mature, which might result in a higher level of performance. On the other hand, a low level of Lean maturity implies the opposite situation in which the implementation of Lean is not recognizable for teams in terms of performance. First, a within case analysis is performed to see how engagement can be defined in this setting. Thereafter, the teams within the same hospital will be compared to each other in the cross case analysis. However, we will also compare the two role model teams of the two

Table 4: Duration level of different teams

(14)

14

different hospitals in order to see how they differ from each other but also to find similarit ies between both teams.

Data collection

In this research, two data collection methods were used: semi-structured interviews and document research. The semi-structured interviews were used to collect data about Lean maturity, second-order problem solving, engagement and performance. The interviews were held with each team and that team consisted of the team leader and a nurse of the core-team, and one nurse that was not from the core-team. The core-team is a special composed team which is responsible for starting the implementation and making non-core team employees enthusiastic to work according the Productive Ward. Since people of the core-team are involved from the beginning of the implementation, the decision is made to interview two of them. Next to that, we interviewed one nurse who was not involved in the core-team in order to avoid a distorted view. Semi-structuring the interviews allows a certain degree of flexibility during the interview and can therefore give more insight in an aspect (Voss et al., 2002). However it will be conducted within boundaries that are established in advance. The interview protocol can be found in Appendix A.

The first set of questions were asked to gather general information about the nurses. This information is used to get insight in the background of the nurses. The goal of the second set of questions was to see whether the nurse performed second-order problem solving. The questions consisted of five scenarios of problems described as similar as possible to their occurrence in the natural setting. These scenarios were developed by Gemmel et al. (2016) and those were used during this research. The scenarios are related to the five broad types of nurse problems as observed by Tucker et al. (2002). Furthermore, second-order problem solving was measured by asking an open question to find out whether nurses could give us an example of second- order problem solving applied by them. The third set of questions were about the engageme nt of employees in order to find out what engagement exactly means in this setting, what it consist of and what the nurses find most important about it. One engagement question was based on the Utrecht Work Engagement Scale (UWES) developed by Schaufeli & Bakker (2002). That question consists six different descriptions of engagement and the nurses had to rank them in order to find out which description is seen as most important. The last set of questions were about Lean maturity. Those questions were based on the enablers and practices of Malmbrandt

& Åhlström (2013). However, the specific questions about enablers and practices were translated to the healthcare setting in order to make it useful for this research. Besides that, not all questions about enablers and practices formulated by Malmbrandt & Åhlström (2013) were asked in the interviews. Some of the enablers were not asked in order to ensure that there was no overlap between the different constructs. Also not all practices were asked to the nurses, because those practices were not relevant when implementing the Productive ward programme.

Therefore, we made a careful selection to measure Lean maturity which resulted in three enablers and three practices. Finally, we asked one question about performance in order to get insight in whether the Productive Ward delivers what it promises, namely releasing more time for direct patient care. Therefore, performance is operationalised in terms of perceived performance which is about the fact whether nurses experienced an improvement of the implementation of Productive Ward programme. The interviews were conducted with mult ip le

(15)

15

investigators and they took on average 45 minutes. According to Eisenhardt (1989), using more investigators can enhance the creative potential of the teams and convergence of observations increases confidence in the findings. All interviews were audiotaped and transcription is done by the same investigator who conducted the interviews. Transcribing all the interviews resulted in 617 pages transcript with in total 297.315 words. Table 5 gives an overview of what was measured during the interviews.

As mentioned earlier, besides the semi-structured interviews we also studied documents.

Studying the documents gives the advantage to access a data source that is not created as a result of the case study (Yin, 2009). For this research, a document concerning the duration of the implementation of Productive Ward was studied. This gives us insight into whether the duration of the implementation plays a role in this research. Besides that, we tried to look into 10-point checklists and Multi Moment Recordings. The 10-point checklist is a measureme nt method to get insight in how well the team is doing a specific module and how far they are with it. However, most checklists were not up to date and consisted many missing data. The Multi Moment Recording is a measurement tool for performance. With those recordings it is possible to see on which activity the nurses spend most time on. However, most teams did not conducted a Multi Moment Recording through which we had missing data. Therefore, both documents will not be used in this research.

Data coding and analysing

After the collection of data and transcribing the interviews, coding was used to categorise the data and to find a pattern or structure in the data. According to Miles and Huberman (1984) coding can be used for data display, conclusion drawing and verification. It is important to code data since it provides the analyst with a formal system to organise the data, uncovering and documenting additional links within and between concepts and experiences described in the data (Bradley et al. 2007). The data coding was done with the use of Atlas.ti. Coding the data can be done in two ways; the deductive way or the inductive way. In this research, the decision is made to use the inductive way of coding for the concept employee engagement. This way of coding provides the opportunity to develop codes based on the data itself (Bradley, 2007). Each interview was reviewed in detail in order to select quotes and assign codes to it that reflect the

Table 5: Overview of what was measured during interviews

(16)

16

concept. The assigned codes were merged to reduce the amount of codes and aggregate dimensions were formulated. This within case analysis allows to see unique patterns of each case before the patterns are generalized. Besides that, the within case analysis helps to become more familiar with the data and helps in reducing the amount of data (Eisenhardt, 1989). Two researchers coded the same data for engagement in order to improve the breadth and depth of the analysis and to make sure that the findings are not biased. The full coding tree can be found in Appendix B.

In order to analyse the constructs Lean maturity, second-order problem solving and performance we decided to make specific levels for them. Those levels were developed based on the theoretical background sections and can be found in Appendix C. Each nurse was assigned to a specific level. In order to make sure that we placed the nurse in the right level, each interview was read by two researchers. Together they made sure that they were unanimo us to put a nurse in a specific level.

After reducing the amount of data by coding, the concept Lean maturity still existed of three enablers and three practices. In order to identify, check and verify for an underlying structure of maturity, a factor analysis was performed. Performing this factor analysis showed that a new scale could be developed since all factors loaded on one dimension. This unidimensionality is showed in Table 6. Besides the factor analysis, the reliability of the scale was also tested. This was done by a measurement that indicates the degree to which the items that are supposed to measure one underlying concept are consistent with each other. The threshold for Cronbach’s Alpha is above 0.7. We measured Lean maturit y by means of 6 items, three enablers and three practices, all on a scale from 1 to 5. As a result of the test we found that the scale was highly reliable (α = 0.85).

In order to validate the construct Lean maturity and to check for manipulation, a correlation test was performed. The manipulation check is meant to check whether the correlation between two variables is high enough to say that they measure more of less the same. The correlation between duration and Lean maturity shows a significant positive relationship (r = 0.57, p < 0.001). The correlation between both variables is strong enough. Apparently, not just duration but also other factors determine the level of Lean maturity.

After completing the within case analysis and statistical tests, the cross case analysis was performed. This is done by following the steps from Eisenhardt (1989) as the categories that were identified in each team were carefully compared to find differences and similarit ies between the different cases. However, before starting the cross case analysis we had to categorise the teams in a level of Lean maturity and second-order problem solving varying from low to high. Although the nurses were already assigned to a specific level we had to reschedule them, since there was a great variance within a team. This variance can be seen in the figures in Appendix D. The variance in a team can be declared through the fact that the three differe nt

Table 6: Factor loadings Lean maturity

(17)

17

nurses within one

team perform

different functio ns.

Due to this variance it is not possible to average the scores of the three team members. Therefore, we made a specific ranking for each function based on the frequency of the scores in order to compute the level to which they belong.

These rankings can be found in Table 7. The table showed for example that the nurses that were not involved in the core team are already in a high mature setting when they scored above 2.5. This is in contrast to the team leader which had to score above 3.0 to come in the high mature level. However, as can be seen in the table, we did not make a distinction in the level of second-order problem solving for the different functions within a team. This is through the fact that below three a nurse did not perform second-order problem solving which is for each nurse the same regardless their function. Therefore, we decided to rank the level of second-order problem solving the same for each nurse. That resulted in a specific level for each nurse which can be found in Appendix E. However, we still had to make one level of the three nurses.

Therefore, we decided to look at each team in depth to see how they scored. When the nurses scored separately high on maturity that team was automatically placed in a high maturity level.

Nevertheless, there were some teams that still scored differently on for example maturity (MLL or HLL). For those teams we decided to place them in the category in which most of the team members were. So MLL and HLL teams were placed in a low level of maturity. Finishing the categorization resulted in seven different levels which vary in maturity level and second-order problem solving level. Table 8 shows an overview of the selected levels and the amount of teams in that specific level.

Table 8: Levels for cross case analysis Table 7: levels for rescheduling the nurses

(18)

18

Results

In this section the results of the quantitative analysis and the qualitative analysis are presented.

The quantitative analysis gives us more insights in whether there is a statistic relations hip between the constructs Lean maturity, second-order problem solving and performance. The qualitative analysis is divided into the within case analysis and the cross case analysis. The within case analysis will show how engagement is defined by nurses which is presented in the coding scheme. However, this within case analysis says nothing about the moderating role of engagement in the relationship between Lean maturity and second-order problem solving. In order to investigate the moderating role of engagement a cross case analysis is performed . Several levels that show differences or similarities are compared to each other to make sure that the role of engagement can be explored.

Background statistics

In total, 45 nurses participated in this research. Those nurses are being part of one of the 15 teams of which approximately 84% was female and 16% was male. Besides that, 15 nurses were team leaders, 15 of them were involved in the core team and the remaining 15 nurses were nurses outside the core team. 71% of the nurses worked less than 10 years in the hospital.

However, 29% worked longer than 10 years in one of the hospitals. An overview of the descriptives can be found in Table 9.

Gender Function Years of

Employment

Female 38 Team leader 15 0-5 16

Male 7 Core team 15 5-10 16

Total 45 Not core team 15 10-15 5

Total 45 15-20 1

20 or more 7

Total 45

Table 9: Descriptives

Correlation analysis

In order to explore whether there is a significant relationship between the dependent and independent variables we make use of the correlation analysis in SPSS. A correlation tells us something about the extent to which some variables are related to each other. In order to find out whether Lean maturity, second-order problem solving and performance are related to each other we performed a correlation analysis. The results of that analysis are showed in Table 10.

Table 10: Correlation between variables

The relationship between Lean maturity and second-order problem solving shows a clear significant positive relation (r = 0.72, p < 0.001). This means that an increase in the level of

(19)

19

maturity leads to an increase in the level of second-order problem solving. Besides that, the correlation analysis also showed a significant direct relationship between Lean maturity and performance (r = 0.72, p < 0.001). Finally, it can be derived from the table that second-order problem solving positively and significantly influence s performance (r = 0.70, p < 0.001).

Regression analysis

After performing the correlation analysis to see whether there is a statistical relations hip between the variables, it is important to figure out how they are related to each other. This is important since one of the research questions stated that Lean maturity will influence second- order problem solving which in its turn will influence performance. This means that the level of second-order problem solving acts as a mediator since the assumption is that the potential impact of Lean maturity on performance is driven by the presence of second-order problem solving. Therefore, the regression analysis was used to explore the mediating role of second- order problem solving. According to Baron & Kenny (1986) the following conditions must be met in order to establish mediation. In the first condition, the independent variable must affect the mediator. As we can see in Table 11 Lean maturity, which is the independent variable affects the mediator second-order problem solving significantly, β = 1.79, t = 6,80, p < 0.00. This means that the first condition is accomplished. The second condition stated that the independent variable must be shown to affect the dependent variable. This is also the case in our regression analysis because Lean maturity significantly influences performance, β = 0.92, t = 6.70, p <

0.00. The third and final condition specified that the mediator must affect the dependent variable. Looking at Table 11 we can see that second-order problem solving significa nt ly influences performance, β = 0.20, t = 2.69, p < 0.05. All conditions are met in this case, which means that there is mediation. There is perfect mediation when Lean maturity has no effect when controlling for the mediator second-order problem solving. However, in this analysis we can see that Lean maturity still influences performance while controlling for second-order problem solving. Therefore, the relationship between Lean maturity and performance is not fully mediated but partially mediated. This means that the direct effect of Lean maturity on performance decreases, however not to zero, when controlling for second-order problem solving.

Table 11: Regression analysis

(20)

20

Besides the regression analysis, you can also use the Sobel Test to measure whether the mediator significantly carries the influence of an independent variable to a dependent variable.

This test was performed as an additional support for the regression analysis. It is a specialized t-test that provides a method to determine whether the reduction in the effect of the independent variable, after including the mediator, is a significant reduction and therefore whether the mediation effect is statistically significant. The Sobel test shows in this research that mediatio n through second-order problem solving is significant (Z = 2.64, p < 0.01).

Within case analysis

In this section the results of the within case are presented. This analysis will give us more insight in how engagement is defined by nurses when implementing the Productive Ward programme.

As a result of the within case analysis, employee engagement seems to consist of the attributes;

collaboration, ambition, dedication and vehemence. These attributes are developed through inductive coding and the concepts that belong to an attribute should fit the definition of the attributes. The attribute collaboration entails all concepts that focus on working with the whole team whereas ambition is more about the focus of the team towards a goal. Dedication is about the feeling that the nurses get from their work whereas vehemence is more about the intens it y with which they work. Each attribute will be elaborated in depth by giving quotes from the nurses. These quotes refer to the number of the team followed by the member of that team. The teams will vary from 1 to 15 and the members can be the team leader which has number 1, the core team member which has number 2 and finally the nurse who is not involved in the core team is referred to as number 3.

Collaboration – collaboration is one of the attributes of engagement and consisted of the concepts participation, thinking along and showing interest. One of the nurses (15.3) stated that

“I think that engagement already indicates that we are doing something together. You have to find solutions together before you can make a change.” This statement highly relates to the collaboration of the whole team. Another nurse said about participation (5.1); “That it is possible to ask somebody for a favour without causing conflict. Besides that, it is important that everyone helps in thinking about how we can improve the processes here on the department.”

Based on this, thinking along is also something that falls within the attribute collaboratio n.

Additionally, the nurse from team 12 (12.3) said that; “engagement is, when I take a look at myself, that you just think along with the things that take place in the department.” Finally, a nurse quoted (4.3): “Engagement to the developments on the department, protocols. That you are aware of it. Well, actually showing interest.” These statements shows that collaboratio n can be seen as a part of engagement. When you have more collaboration within the team, the engagement level will be higher.

Ambition – ambition is the next attribute of engagement. Asking nurses about what they think engagement entails, they focuses on three aspects that fall within the attribute ambition. The first one is the focus on patients. One of the nurses said; “Engagement is that everyone wants a high level of quality of care.” Someone else stated that (3.1); “Engagement is definitely that you take time for patients. I think that is absolutely a priority. You have to do your actions with sufficient knowledge of a patient and you can’t do something of which you don’t have enough knowledge.” Besides the focus on patients, continuous improvement is also seen as an ambition

(21)

21

by nurses; “Engagement is that you’re doing your work in the best possible way and that you also contribute in improving it” (5.1). Finally, the responsibility that nurses get seems to play a role; “Engagement is for me how much sense of responsibility you have. Through Productive Ward everyone gets more responsibilities. I am not the only one responsible for the things that happen here on the department.” (11.1). By increasing the level of responsibility, the nurses are becoming more aware of the goals of the team. So the greater the ambition of the members, the more engagement you could expect.

Dedication – dedication says something about the enthusiasm about your work, the appeasement that you get from it and how you are challenged by your work. For example, the team leader from team 9 (9.1) stated; “It is really important that people are enthusiastic about the things they are doing here.” Additionally, another team leader (4.1) said that; “The most important thing is that the nurses are enthusiastic.” Appeasement is another element of dedication and is mentioned by a nurse from a core team (13.3) as; “I work here for my satisfaction. My heart is lying here. I never want to do something else than taking care for patients.” Finally, different nurses said that they are seeking for a challenge in their work:

“Well, I think that the most important thing is that my work is useful and challenging. I need to be inspired through my work.” (3.2). So showing more dedication in your work, seems to improve the level of engagement.

Vehemence – vehemence is more about the passion and pleasure that you have in your work and how motivated you are. Passion is mentioned by several nurses as; “Caring about the business” and “I love to do my work.” Pleasure is also often mentioned by nurses;

“Engagement is that someone is coming with pleasure” and “That you have fun in your work, but you also have to make it pleasant for others.” Passion and pleasure are both seen as important factors by nurses when asking to define engagement. Additionally, also the motivation of the nurses is important. Motivation is described by a team leader as (6.1); “That someone does not only do the work because he gets payed, but he also must be internally motivated.” So, when you have more passion and pleasure in your work and you are more motivated to do your work, the engagement seems to be higher.

Table 12 gives an overview of the different attributes of engagement and the corresponding concepts which are described in the previous paragraphs. Collaboration is about involve me nt of the whole team and doing something together. Ambition focuses on the goals of the team whereas dedication is more about deriving significance from your work. Finally, vehemence is more about the intensity you work with. All four attributes seems to play a role in the engagement of nurses. However, how engagement works as a moderator on the relations hip between Lean maturity and second-order problem solving will be presented in the cross case analysis.

Table 12: attributes and corresponding concepts of engagement

(22)

22 Analysis of rankings – defining engagement

Besides the fact that we tried to develop a definition of engagement, we also asked nurses during the interviews to rank six descriptions of engagement which are in line with the definition of engagement of Schaufeli & Bakker (2004). The six descriptions – want to work, enthusias t ic about work, feeling that time flies, work is useful and challenging, energy from your work and complete focus on work – could in its turn be divided into three attributes namely vigour,

dedication and

absorption. The scores were averaged in order to find out which of the attributes is seen as most important by nurses. The results of the rankings are displayed in Figure 3. It is important to know that the higher the score, the better the attribute is ranked.

As can be derived from the figure, dedication is seen as the most important attribute according to the nurses and scored on average a 4.5. Absorption in contrast scored the lowest with an average of at most 2. Vigour scored on average a 3.9 which is also quite good. Based on this, absorption seems to be less important compared to the other two attributes. This result will be used for showing which attributes of engagement are seen as important by nurses and how it relates to the definition of engagement defined in the previous section. What each of the six descriptions individually scored can be found in Appendix F.

Cross case analysis

In this part the results of the cross case analysis will be shown. As mentioned earlier, the teams had shown differences in the level of Lean maturity and second-order problem solving. That resulted in a specific pattern which is presented in Figure 4. Besides the level of Lean maturit y and second-order problem solving, the level of duration is also showed in this figure. When we look at this figure it can be seen that generally when there is a low level of maturity, there is also a low level of second-order problem solving. This also applies in the other way around so when there is a high level of maturity there will also be a high level of second-order problem solving. However, there are some teams that did not perform as you should expect (team 11, 4, 3, 8 and 2). Therefore, five comparisons will be made in order to see how engagement plays a role in the relationship between Lean maturity and second-order problem solving. Again we will make use of quotes to show whether teams are engaged or not. We will refer to the quote s as we also did at the within case analysis.

Figure 3: ranking engagement

(23)

23

Cross case comparison 1 – in the first comparison the most extreme cases are compared to each other. This means that the teams that scored low on maturity and second-order problem solving are compared to the teams that scored high on maturity and second-order problem solving.

When comparing both levels, we can see that there is a difference in level of engagement based on the quotes of the nurses. First, it has to be mentioned that there is a difference at the start of the implementation. The teams that scored high on the two variables are both pilot departments.

This means that the whole team is excited to be the first team in the hospital in which Productive Ward is implemented and there is also involvement of higher management in order to see whether it is a successful implementation. Besides that, when comparing both levels we see that there is a difference in the level of collaboration. In the high scoring teams the whole team is engaged whereas in the low scoring teams the nurses had no idea what the core team is doing.

Additionally, the high scoring teams are more enthusiastic and motivated to work according the principles of Productive Ward compared to the lower level in which they said that they are not motivated to take for example place in the core team. Finally, working towards the same goals seems to make a difference in the level of engagement. Based on this comparison we could say that engagement plays a role in the relationship between Lean maturity and second-order problem solving. The more collaboration, ambition, dedication and vehemence, the more engaged nurses are in the implementation.

Figure 4: Pattern cross case analysis

Hospital B

12 14 9 6 11 4 3 10 5 7 13 8 2 1 15

Second Order Problem Solving Level

Hospital A

Duration [Months]

6-12

< 12

Maturity Level H

M

L

H

M

L

> 24

18-24

12-18

(24)

24

Cross case comparison 2 – in the second comparison the two teams that scored high on maturit y and second-order are compared to each other because they both are pilot teams from two different hospitals. However, the duration of the teams differ from each other, team 1 worked longer on the implementation than team 15, which makes this an interesting comparison.

Looking at the quotes, we can see that there are similarities between the teams rather than differences. Both teams show that a degree of participation and knowing what is going on has ensured that the nurse showed more engagement. Besides that, also working towards the same goals and having responsibilities seems to influence the level of second-order problem solving.

Furthermore, the teams pointed out that a degree of enthusiasm and motivation provides higher levels. This shows that the four categories of engagement are crucial in the relationship between Lean maturity and second-order problem solving.

Figure 5: cross case comparison 1

Figure 6: cross case comparison 2

Referenties

GERELATEERDE DOCUMENTEN

De vangsten zijn berekend voor de bordentrawlvisserij voor 16 en voor de garnalenvisserij voor 6 soorten welke in de vangstdatabase gespecificeerd konden worden binnen de twee ICES

A model of propagating rea tion fronts is given for simple auto atalyti.. rea tions and the stability of the propagating rea tion fronts

Therefore, under high workload employees are less likely to show citizenship behavior as a result of high LMX relationships, because these employees do not have the mental or

And certain OC dimensions were found to be positively associated with LM extent, including future orientation and uncertainty avoidance for both lean soft and hard

Appreciative Inquiry Work engagement Relatedness Perceived expertise affirmation Self-efficacy Direct effect c’ a1 a2 a3 b1 b2 b3 Appreciative Inquiry Work engagement

collect water meter readings and send the information to the gateway [58]. They also developed a WaterGrid-Sense, which is a smart interface platform with the ability to monitor

Although proficiencies in languages and literacies are often included in studies of academic achievement of South African students as a contributing factor, we could find

Murray, Identification of clinical isolates of anaerobic bacteria using matrix-assisted laser desorption ionization-time of flight mass spectrometry, Eur. Burnham, Optimization of