An overall desired goal does not make the change -

(1)

An overall desired goal does not make the change

-

A multilevel study on the influence of team climate, functional

uncertainty, implementation interventions and readiness for change on the

adoption of change in healthcare settings

October 22nd, 2012

Master thesis, MSc BA, specialization Change Management University of Groningen, Faculty of Management and Organization

SUSAN LAURA WESSELINK Student number: 1741365 van Iddekingeweg 138-53

9721 CL Groningen wesselink.sl@gmail.com

Supervisor / university M.A.G. van Offenbeek Supervisors / field of study

G.A. Welker A.C. Hobo

(2)

2

Introduction

A great number of change processes fail due to many varying circumstances (Burnes, 2004a) and many factors contribute to the effectiveness of implementation of change (Armenakis, Harris and Mossholder, 1993). One of those factors contributing to the effectiveness of change implementation is the readiness for change experienced by people involved in the implementation. According to Armenakis et al. (1993) and Jones, Jimmieson & Griffiths (2005), UHDGLQHVVIRUFKDQJHFDQEHGHILQHGDVµWKHSRVLWLYHIHHOLQJRIHPSOR\HHVWRZDUGV FKDQJH¶&KDQJHUHDGLQHVVLVFORVHO\UHODWHGWRWKHµXQIUHH]LQJ¶VWDJHLQ/HZLQ¶VWKUHH-stage model to change (Lewin, 1951 in Burnes, 2004b). These three stages do not guarantee successful change, but constitute the three main activities that should be managed in change (Kanter, Stein & Jick, 1994). Unfreezing is the disconfirmation of the validity of the status quo, the induction of guilt or survival anxiety and creation of psychological safety (Burnes, 2004a). Employees must have the feeling that the change will influence them and the

organization in a positive way. Furthermore, they must recognize the need for change. A low readiness for change causes lower willingness of employees to cooperate in implementing change and can therefore seriously delay implementation or can even lead to employees abandoning the change (Rafferty & Simons, 2006). Next to the unfreezing stage, Lewin (1951) described two additional main activities that should be managed in order to make change happen. After unfreezing the employees of an organization, the next stage is moving, in which the organization transforms from old to new guidelines, routines and/or actions. The last stagHLQ/HZLQ¶VPRGHOLVUHIUHH]LQJ5HIUHH]LQJDLPVDWVWDELOL]LQJWKHRUJDQL]DWLRQDW the new equilibrium. This activity is important in order to ensure that new behaviours are relatively safe from regressions and the change will be sustained, (Burnes, 2004a).

There are various models that provide guidelines on how to implement change. Elements of the models are focused at the unfreezing, moving and refreezing stages as described above. All models have considerable consensus about important aspects that should be managed in change. Kotter (1995) suggests eight steps in order to change successfully. Another often-used model for successful change programs is that of Kanter, Stein & Jick (1992). They GHVFULEHµ7HQ&RPPDQGPHQWV¶IRUH[HFXWLQJFKDQJH7KHVHWZRmodels together provide detailed guidance for implementing change (Burnes, 2009).

Despite these seemingly clear prescriptive models for how to change, implementing change successfully still tends to be difficult for organizations. One reason for this may be that managers do not wish to follow the clear guidelines (Richardson & Denton, 1998), or that the models are only seemingly clear. Another reason is that more specific influencing contextual factors are present in certain situations (Kotter & Schlesinger, 2008). Almost all of the models only look at broad characteristics that should be managed whenever you are involved in organizational change, neglecting contingencies. However, according to Kotter & Schlesinger (2008), successful change efforts seem to be those where choices are both internally

(4)

4 organizational context and what works in one FRQWH[WPD\IDLOLQDQRWKHU´ This raises the question which contingency factors are present.

The influence of the contingencies on the implementation process of a quality instrument in a healthcare setting will be studied. A wide variety of quality improvement (QI) instruments can be used in healthcare and research shows that effectiveness varies between settings (Hulscher, Laurant & Grol, 2003). It is thus interesting to study specific situational factors, in order to increase effectiveness of interventions in the healthcare sector. This study focuses on two contingencies: team climate and functional uncertainty. They will be assessed in a healthcare institute.

In hospitals, one of the situational factors that can be of influence is team climate. Teamwork is essential in hospitals and team effectiveness is of influence on the success of the change implementation. Team effectiveness in turn is dependent on the team climate (Bower, Campbell, Bojke & Sibbald, 2002). According to Heineman (2010) differences between the specialties in hospitals are very large and not every profession is focused on working together for the collective good. Every department in a hospital has a different composition and team climate can differ; this might mean that for one team it is easier to adapt to change than for other teams.

Another contingency that might be interesting to look at is functional uncertainty. Functional uncertainty is one of the five situational domains in implementation described in the

RABSODY instrument of Van Offenbeek, (1993). It contributes to the complexity of the implementation. Functional uncertainty focuses on the risk associated with the

implementation of technical innovations, whereby old tasks disappear or change and new tasks are being created. Innovation-system fit is herewith important (Greenhalgh, Macfarlane & Kyriakidou 2004). Furthermore, the risk that the implementation will not be effective is influenced by the impact the change will have and the experience that employees already have with comparable quality instruments (Van Offenbeek & Koopman, 1996).

This study combines two sub-studies on two levels to assess all the variables and to use all the available information of the quality instrument implementation under investigation. The first sub-study is on individual level. The concepts readiness for change and team climate are central in this sub-study, since those are concepts that individuals involved in an

implementation can experience more or less.

The second sub-study is a study on department level, since precise measures of the

(5)

5 Aim of the research

This study has different aims. The first aim is to analyze the influence of two contingencies on the relation between implementation interventions and the successful implementation of a quality instrument implementation. Secondly, the mediating role of readiness for change will be explored. Finally the direct influence of implementation interventions and the ultimate implementation outcome will be analyzed.

This study also aims at providing practical insights. In the end this research should provide healthcare institutions with more knowledge about the influence team climate and functional uncertainty have on the success of implementations. Furthermore, should insight into different dimensions of readiness for change help managers determine specific actions in order to influence those dimensions.

O rganizer

(6)

6

Literature study

In this chapter the central concepts of this study will be introduced as well as the expected relations between the concepts. The relations will be displayed in the conceptual models in the end of the chapter as well as in the corresponding research questions.

2.1 Implementation of change

From research we have learned that although the frequency of change increases, only few implementation projects are as successful as planned time (KPMG, 1997 and Standish Group, 1996 in: Thomas, Delisle, Jugdev and Buckle, 2002). There are many factors contributing to the effectiveness of implementation of change (Armenakis et al., 1993). (PSOR\HHV¶UHDGLQHVVIRUFKDQJHLVRQHRIWKRVHIDFWRUVFRQWULEXWLQJWRWKHVXFFHVVRIFKDQJH since employees can facilitate or hinder the change process (Armenakis et al., 1993). Existing literature provides organizations and managers with implementation models that provide guidelines for successful implementations. In this study, two widely accepted models provide the baseline for investigating the impact of implementation interventions on readiness for change and the adoption to the implementation. The first model is the eight-step model of Kotter (1995) and the second model is the Ten Commandments for executing change of Kanter et al. (1992). These two models are chosen because together they provide detailed guidance for implementing change (Burnes, 2009). First, these two models will be discussed in more detail.

.RWWHU¶VHLJKWVWHSVWRVXFFHVVIXOFKDQJH

In his model, Kotter (1995) described eight steps to transforming an organization as can be seen in the table (1) below. The eight steps are a process and not a checklist. Neglecting just one of the stages can lead to serious problems in implementation of change (Burnes, 2004). .RWWHU¶VPRGHOLVEDVHGRQUHVHDUFKLQDZLGHUDQJHRIFDVHVWXGLHVRIDURXQG$PHULFDQ organizations. This resulted in a large database that was used to generate a normative approach to change (Buchanan, Fitzgerald, Ketley, Gollop, Jones, Saint Lamont, Neath & Whitby, 2005).

Kanter et al. (1992) Ten Commandments to executing change

Kanter et al. (1992) have a somewhat more detailed model, wherein they add two other influencing factors, supplementing the model of Kotter (1995). The model of Kanter et al. (1992) is based on a review of change tactics that literature provided around that time. The Ten Commandments together are the essence of the most mentioned advices on how to execute change (Kanter et al., 1994).

(7)

7

Table 1. Implementation models of Kotter (1995) and Kanter et al. (1992).

Kotter (1995)

Eight steps to successful change

Kanter et al. (1992)

Ten Commandments to executing change

Create a sense of urgency Analyse the organization and its need for change

Form a guiding coalition Create a shared vision and a common direction

Create a vision Separate from the past

Communicate the vision Create a sense of urgency Empower others to act on the vision Support a strong leader role

Create quick wins Line up political sponsorship

Build on the change Craft an implementation plan

Institutionalize change Develop enabling structures

Communicate, involve people and be honest Reinforce and institutionalize the change

2.2 Readiness for change

In this paragraph the importance of readiness for change will be described as well as the proposed relations between implementation interventions and readiness for change and readiness for change and the adoption to the change. Furthermore, the definition of readiness for change for this study will be given.

The two models described above share one central aim. With the first steps, they explicitly try to influence the level of readiness for change of employees. Readiness for change can be defined as the positive feeling of employees towards change such that it will influence them and the organization in a positive way and that they recognise the need for change (Armenakis et al., 1993; Jones et al. 2005). Readiness is one of the most important factors influencing HPSOR\HHV¶initial support for change initiatives (Armenakis et al., 1993). Influencing the amount of readiness is therefore important, since the opposite is also true; low levels of readiness for change causes lower willingness of employees to cooperate in

implementing change and can seriously delay implementation or can even lead to employees abandoning the change (Rafferty & Simons, 2006).

As described in the introduction, readiness for change is often associated with the unfreezing VWDJHRI/HZLQ¶VWKUHH-stage model. Lewin (1951) believed that the stability of human behaviour was based on a quasi-stationary equilibrium supported by a complex field of driving and restraining forces. Before new behaviour can be adopted successfully, the equilibrium needs to be destabilised. According to Schein (1979) many change efforts run into resistance or failure due to not providing an effective unfreezing process. Armenakis et DOVWDWHWKDW³FUHDWLQJUHDGLQHVVIRUFKDQJHLQYROYHVSURDFWLYHDWWHPSWVE\DFKDQJH agent to influence the beliefs, attitudes, intentions and ultimately the behavior of a change WDUJHW´

(8)

8 readiness for change. The first dimension is emotional readiness for change, which captures the feelings about a specific change project. Secondly, cognitive readiness for change focuses RQWKH³EHOLHIVDQGWKRXJKWVRUJDQL]DWLRQDOPHPEHUVKROGDERXWWKHRXWFRPHVRIWKHFKDQJH´ The last dimension is intentional readiness for changeZKLFKUHSUHVHQWV³WKHH[WHQWWRZKLFK HPSOR\HHVDUHZLOOLQJWRSXWHQHUJ\DQGHIIRUWLQWRWKHFKDQJHSURFHVV´%RXFNHQRRJKHHWDO 2009). According to Bouckenooge et al. (2009) ³the multifaceted view of readiness for

change as a triadic attitude instead of unifaceted operationalization is better at capturing the complexity of the phenomenon´. Additionally, this multifaceted construct involves both readiness for change in general as the readiness for a specific change. This means that information will be generated both for this implementation and for future implementation projects (Bouckenooghe et al., 2009:561).

The first steps of the models of Kotter (1995) and Kanter et al. (1992) provide guidance in attempts of change agents to influence readiness for change. Because of the expected influence of readiness for change on the ultimate success of an implementation, readiness for change is expected to have a mediating role between the implementation interventions as described in the models of Kotter (1995) and Kanter et al. (1992) and the implementation outcome.

2.3 Contingency factors

Beyond these models, that provide a guideline for how to execute change, and the importance of readiness for change, various studies have studied the influence of contextual factors on change implementation (Bouckenhooghe et al., 2009). The contingency approach focuses on a fit between the organization and its context, in order for the organization to be effective (Van Offenbeek & Koopman, 1996). Grol and Wensing (2006) concluded that it is often not successful to copy a change intervention that was successful in one organization into another organization. Change needs to be tailored to the needs of a situation in order to achieve the best outcome (Redfern & Christian, 2003). In this thesis, the influence of two contingencies will be studied. The first is the team climate and the second is the functional uncertainty that is present within the change system.

2.3.1 Team climate

In the past years, healthcare changed from individual consultation towards multi-disciplinary teamwork of healthcare practitioners. Optimal collaboration between professionals is

important for offering continuous and coordinated care (Ouwens, Marres, Hermens, Hulscher, van den Hoogen & Wollersheim, 2007).

$FFRUGLQJWR%RZHUHWDOµWhe potential advantages of working in integrated teams in primary care are threefold and involve increases in (1) task effectiveness (improving patient health and satisfaction with care); (2) mental health (the morale and well-being of team members); and (3) team viability (the degree to which DWHDPZLOOIXQFWLRQRYHUWLPH¶

(9)

9 hospitals like the individualistic attitude of doctors (Heineman, 2010), the varying

background of team members and perceived inequality in status (Poulton & West, 1999), team climate is proposed to differ also between teams within one health care organization, because of the highly varying compositions of teams and individual attributes that are contributing to team climate.

The most studied model of team climate is the four-factor model of West (1990) known as the Team Climate Inventory. In Anderson & West (1998) the four different elements of their team climate concept are described as follows:

1. Shared vision and objectives

A vision is an idea of a valued outcome which represents a higher order goal and a motivating force at work. Teams with clear objectives and focus on the direction are more likely to develop goal-appropriate methods of working (Anderson & West 1998). According to

$QGHUVRQ :HVWµVKDUHGQHVVPHDQVthat the vision gains widespread acceptance E\LQGLYLGXDOVZLWKLQWKHWHDP¶.

2. Participative safety

Participative safety exists in situations in which involvement in decision-making is motivated and occurs in a non-threatening environment (Bower et al., 2003). According to West (1990; in Anderson & West, 1998) the more people are able to participate in decision-making by for example having influence, interacting and sharing information, the more likely they are to invest in the outcomes of those decisions. (GPRQGVRQ¶VFRQFHSWRISV\FKRORJLFDO safety is important in this element of team climate. Psychological safety is the shared feeling by team members about how safe it is to take interpersonal risks. Psychological safety can lead to team learning, a process of reflection and interaction, in which team members actively share knowledge and information to contribute to the team functioning (Edmondson, 1999). To make this possible, it is important that the team shares the conception that it is safe to express opinions and that members are not being punished or rejected for making mistakes or asking for help (Edmondson, 1999).

3. Task orientation / Commitment to excellence

This involves a shared concern with quality of task performance, in relation to shared vision or outcomes. This can be characterized by evaluations, modifications, control systems and critical appraisals (West, 1990). This factor also focuses on a climate that supports the adoption of improvements to established policies, procedures and methods (Anderson & West, 1998).

4. Support for innovation

(10)

10 interventions and readiness for change. It is proposed that high scores on the above three factors has a positive influence on the relationship between implementation interventions and readiness for change. This means that when team climate is perceived as high (positive), the unfreezing LPSOHPHQWDWLRQLQWHUYHQWLRQVOHDGWRµPRUH¶UHDGLQHVVIRUFKDQJHDQHPSOR\HH perceives. Team climate can thus be seen as a moderator on this relation.

2.3.2 Functional uncertainty

Whenever change consists of technical innovations, there are certain risks present in the implementation. The concept of functional uncertainty is defined as a risk factor in

implementations within the RABSODY instrument (Jenner et al, 2008). This instrument is developed to diagnose and determine risk factors in implementation of technical innovations within the healthcare sector (Jenner et al., 2008). In this model, risks are described as factors that endanger a successful implementation in terms of sustainable adaptation. Since in the implementation under investigation consists of a technical innovation in the information system used within the surgical trajectory, it is interesting to look at the influence of the risk associated with this change.

Functional uncertainty focuses on the risk associated with the implementation of a system, whereby old tasks disappear or change and new tasks may be created (Van Offenbeek & Koopman, 1996). Functional uncertainty consists of two categories, which together comprise information about the existing situation and the change situation. Information about the existing situation consists of the complexity, stability and knowledge about current tasks. The change situation focuses on the knowledge about solutions for problems, the awareness of goals and needs of the target group, the degree of change within existing processes and the amount of experience with similar technical innovations, in this study quality improvement instruments, that users already have (Van Offenbeek & Koopman, 1996; Jenner et al., 2008). Functional uncertainty is an important variable to determine in implementations, since it defines innovation-system fit (Greenhalgh et al., 2004). Innovation-system fit describes the match between the new system and the organization it is implemented in. This fit can be defined in terms of fit with the strategy, goals, ways of working and supporting technology. A fit between the innovation and the existing system will positively influence success of the implementation (Greenhalgh et al., 2004). The higher the risk of functional uncertainty, the higher the chance an innovation will be unsuccessful. Interventions in the risk domains can influence the success of an innovation by increasing the innovation-system fit (Greenhalgh et al., 2004).

(11)

11 2.4 Implementation outcome

The implementation outcome refers to the success of the implementation is terms of adoption. It is related to more popular terms like institutionalization or sustainability of change.

%XFKDQDQHWDOVD\WKDWµVXVWDLQDELOLW\LPSOLHVWKDWQHZZRUNLQJPHWKRGVDQG SHUIRUPDQFHOHYHOVSHUVLVWIRUDSHULRGDSSURSULDWHIRUWKHVHWWLQJ¶It thus goes further than just acting according to the change. Routines and norms must also be transformed; otherwise the behavior will not be sustained (Burnes, 2009). The last two steps of the models Kotter (1996) and Kanter (1992) are focused on the institutionalization of the implementation. The relationships derived from the literature discussed above, are displayed in the two conceptual models below.

2.5 Conceptual models

+

+ +

Figure 1. Conceptual model for study on individual level

The research questions that represent the first sub-study on individual level are: Does team climate have a moderating effect on the relationship between the

µXQIUHH]LQJ¶LPSOHPHQWDWLRQLQWHUYHQWLRQVDQGUHDGLQHVVIRUFKDQJH" 'RHVUHDGLQHVVIRUFKDQJHPHGLDWHWKHUHODWLRQEHWZHHQWKHµXQIUHH]LQJ¶

(12)

12 -

+

Figure 2. Conceptual model for department level study

The research question under investigation in the second sub-study is:

How does the amount of functional uncertainty influence the implementation

interventions, the adoption of the quality instrument (SURPASS) in usage percentage and the relationship between those two?

(13)

13

Research method

3.1 Implementation context: Surgical patient safety in the U M C G

µ+RVSLWDOVDUHQRWWKHVDIHSODFHVZHZRXOGOLNHWKHPWREH¶$V\VWHPDWLFUHYLHZKDVVKRZQ that 1 in every 150 patients admitted to a hospital dies as a consequence of an adverse event and that almost two thirds of in-KRVSLWDOHYHQWVDUHDVVRFLDWHGZLWKVXUJLFDOFDUH¶ (De Vries, Ramrattan, Smorenburg, Gouma & Boermeester, 2008). The quality of medical care needs continuous improvement, so that medical errors can be diminished. Due to the amount of medical errors that have been reported over the last years, the spectorate (Inspectie Gezondheids Zorg, IGZ) sharpened their guidelines and procedures for medical care and especially for the surgical trajectory (TOP ± toezicht operatief proces, www.igz.nl).

In order to increase patient safety in the surgical trajectory, the Academic Medical Centre in Amsterdam (AMC) has developed a multidisciplinary checklist procedure called SURPASS (SURgical Patient Safety System, www.surpass-checklist.nl), which runs throughout the surgical trajectory and focuses on responsibilities and specific checks that need to be done before the patient can move to the next phase of the trajectory, such as patient identification, information about medication and available resources in the surgery room. Due to this checklist, avoidable complications and mistakes throughout the surgical trajectory were reduced in the AMC. Furthermore, the checklist could also improve cooperation and communication between professionals (De Vries et al., 2010). According to Verdaasdonk, Stassen, Widhiasmara & Dankelman (2009) checklists are often used as a means to improve team coordination. Moreover, they say checklists can be used as a means to create a culture of quality, to support quality control, to assure quality through the standardisation of behaviour as a defence strategy to prevent human errors and/or as a memory aid to enhance

performance.

Due to the enormous success of the checklist developed at the AMC, research has been conducted and empirical evidence exists that the checklist can considerably help in

substantially diminishing the number of medical errors in the surgical. The number of medical errors decreased, due to this procedure, with almost 33% (De Vries et al., 2010). The IGZ handed out the ZorgVeiligPrijs for SURPASS to the AMC, because of the scientific and practical innovation they created (IGZ.nl).

The success of SURPASS for the AMC inspired the UMCG to also implement SURPASS. In September 2010, the paper version of the checklist was bought, because the IT systems were not able to support the digital version of the checklist. This first initiation of the process was not successful and caused a lot of resistance among key stakeholders (anonymous, field of study). Alternative checklists were analyzed to look for possible solutions for the IT problem. However, SURPASS was still found to be the best. This time the digital checklist was

(14)

14 lessons could be learned from previous phases in the process.

3.2 Research design

This study was designed in accordance with the embedded multilevel single case-study (Yin, 2009). The organizational context provided a good opportunity to explore the models. A single case-study design is appropriate for this research, since its aim is to provide information about specific contingency factors influencing the effectiveness of QI

implementations in the UMCG and more general in health care institutes. The study is partly retrospective and thus relies for a large amount on emplo\HHV¶memories.

The study combines two sub-studies, one on individual level and one on department level. For the two sub-studies, the sample was the same. The sample of respondents included medical specialists, medical specialists in training, nurses, nurse practitioners, planners and surgery-assistants. The quality instrument implementation under investigation (SURPASS) belonged WRWKHUHVSRQGHQWV¶GDLO\ZRUNDFWLYLWLHV&DVHVWXGLHVLQJHQHUDOODUJHO\UHO\RQTXDOLWDWLYH data. However, in this case study, quantitative and qualitative data are used, since the combination of data types can be highly synergistic (Eisenhardt, 1989). In this case study, theoretical sampling (Yin, 2009) was done to find whether the largest contrast in the results could be explained theoretically in the way they were expected. Six departments were used to find differences in the variables of the conceptual model. Two departments that were already performing quite well were chosen, as well as two departments that are currently

underperforming. The other two departments were chosen randomly. The departments were chosen from monthly reports of usage percentages.

3.2.1 Research design for individual level sub-study.

For the sub-study on individual level, the focus lies on the concepts µXQIUHH]H¶

implementation interventions, readiness for change and team climate. The only data source that was used for the individual level sub-study was the questionnaire that was handed out to the employees from participating departments.

3.2.2 Research design department level sub-study

The department level sub-study focuses on exploring the role of functional uncertainty in the quality instrument implementation. Beside the questionnaire, also qualitative data sources were used to explore the influence of the concepts in the conceptual model.

Table 2. Case study sampling logic

(15)

15 3.3 Data collection

The qualitative data used were derived from available documents, conversations with the two implementation coordinators of the SURPASS implementation as well as feedback from a planned feedback/evaluation meeting at the end of this research. The available documents were: the implementation plan, highlight reports from meetings with ambassadors,

presentations about SURPASS and evaluation reports.

The concepts from the conceptual model were derived from the literature. The questionnaire (see appendix A) consisted of questions about: implementation characteristics, team climate, functional uncertainty, readiness for change and implementation outcome. As described below, the questions about functional uncertainty, readiness for change and team climate were based upon validated measures from literature. Questionnaires were handed out in the

participating departments. Employees had three weeks to fill in the questionnaires. After this time, the questionnaires were collected.

3.4 Procedure

Since the selected multi-item scales were all in English, translation to Dutch was needed. Translation of the questionnaires was done personally. However, translation was checked by the supervisors of this study, from both the UMCG and the RUG.

Several people in the UMCG tested the final questionnaire in a pilot: one of the

implementation coordinators, the supervisors of this thesis (RUG & UMCG) and the members of the sounding board. Discussions led to some adjustments in the questionnaires. After the pilot, questionnaires were ready to be handed out in the participating departments.

Questionnaires were handed out in different ways. For medical specialists and assistants, they were provided at their secretary in envelopes with a return envelope enclosed. For wards, questionnaires were delivered in mailboxes and nurses had the opportunity to return the questionnaires in closed return boxes in their coffee rooms. One week after the distribution of the questionnaires a reminder was delivered to each secretary and ward. The form of the reminder was a box of candy with a colored paper attached, on which a short description of the research was written. Two weeks after the distribution, the first questionnaires were collected. Since response percentages stayed behind, another reminder was send. Every department was called and at some departments another short briefing followed. After three weeks the final questionnaires were collected. After the data analyses were conducted, a feedback meeting was planned to communicate and discuss the results with stakeholders in the SURPASS implementation. Each participating employee was invited. Eight employees were present and represented four of the six participating departments.

3.5 Measurement of constructs

In appendix A, the complete questionnaire can be found. Implementation interventions

(16)

16 Contingency factors

The contingency factors are measured in following ways: Team climate

Team climate was measured with the shortened version of the Team Climate Inventory. This is a 14 item questionnaire, which focuses on the four dimensions of team climate according to West (1990). The dutch TCI-14 was developed by Strating & Nieboer (2009). Since the fourth dimension team innovation was not applicable for this study, these items were left out of the questionnaire, leading to an 11-item TCI questionnaire. The answers were measured on a 5-points Likert-like scale.

Functional uncertainty

Functional uncertainty was measured through the (11) questions developed within the RABSODY method of risk identifying in implementation (Offenbeek & Koopman, 1996).

Questions were aimed at defining the amount of risk associated with implementation of the SURPASS checklist in different areas and each is measured on a 3-points scale ranging from low-medium-high. The total amount of the risk type functional uncertainty was the sum of the answers the employees of the respective department gave (formative measure).

Readiness for change

Readiness for change was measured using the validated questionnaire of Bouckenhooghe et al. (2009). They divide readiness for change in three sub categories: emotional, intentional and cognitive readiness for change. The total amount of items is 14. The answers are measured on a 7-points Likert scale.

Implementation outcome

(17)

17

Table 3. Summary of questionnaire

3.6 Data analysis for the individual level sub-study

In this section the validity and reliability of the items in the questionnaire will be described. The construct validity determines whether the constructs measure what they supposed to measure (Cooper & Schindler, 2008). To determine the construct validity, in this thesis factor analyses in combination with Cronbach alpha values are used. The values for the CURQEDFK¶V alpha are acceptable when they are between 0.6 and 0.7, values above 0.8 are good (Sekaran and Bougie, 2010). The results of the analyses will be described below. For the items of the concepts that were newly developed for this research (implementation interventions and implementation outcome) an explorative factor analysis was performed. For the variables that were based on existing validated questionnaires (readiness for change and team climate), a confirmatory factor analysis was performed because these multi-item scales were applied in a new setting. The results of the EFA and CFA can be found in appendix B.

Implementation interventions

To measure the experienced exposure to implementation interventions, the steps of Kotter (1996) and Kanter (1992) were divided into 10 items. A division in unfreezing, moving and refreezing was expected (3). However, the explorative factor analysis showed that items loaded on two different factors, instead of three. Moreover, the exact division between

Variable/ Components

&KURQEDFK¶V

alpha

Level on which items are assessed

Number of items

Implementation interventions Individual and

department level

Unfreeze 0.86 5

Move/Refreeze 0.76 5

Team Climate 0.90 Individual level

11

Functional uncertainty - Department level

Formative measure 10

Readiness for change Individual level

Emotional readiness 0.85 6

Intentional readiness 0.86 3

Cognitive readiness 0.80 5

Implementation outcome Individual level

(18)

18 unfreezing, moving and refreezing was not reproduced. Only the interventions of unfreezing ORDGHGWRJHWKHUZLWKWKHLWHPµXVHRIDFOHDULPSOHPHQWDWLRQSODQ¶RQRQHIDFWRUThis might be explained by the fact that a clear implementation plan is considered to be part of a good preparation of an implementation and thus belongs to the unfreeze items. Moving and

refreezing items both loaded on the same factor, possibly because µUHIUHH]LQJ¶ZDVRQO\DVNHG in one item and the implementation process was still running during this study. This might have been confusing for the participants. The implementation interventions are, therefore, divided into two variables: unfreezing and moving-refreezing.

Team Climate

To measure team climate, the validated questionnaire of Strating & Nieboer (2009) was used. The questionnaire consists of 11 items. The confirmatory factor analysis showed that all items loaded high on one factor as expected.

The different items of the contingency factor functional uncertainty did not load on any factor in the factor analysis. This was expected since the different items do not measure one specific construct, but are formative measures. Formative measures are commonly used for constructs conceived as composites of specific component variables (Edwards & Bagozzi, 2000). Readiness for change

Readiness for change was expected to load on three different factors, since the used questionnaire from Bouckenhooghe et al. (2009) measures emotional, cognitive and

intentional readiness for change. The confirmatory factor analysis revealed that two items of emotional readiness for change did not load on any factor. They could thus have been LQWHUSUHWHGZURQJO\7KH&URQEDFK¶VDOSKDIRUWKHLWHPVZDVĮ =.45. When the two

emotional readLQHVVLWHPVWKDWORDGHGGLIIHUHQWO\ZHUHH[FOXGHGWKHYDOXHRIWKHFURQEDFK¶V alpha for this sub-construct was Į =.84. Only two out of the five items of cognitive readiness for change loaded on the expected factor, two of the other items did not load on any factor and one of the items loaded on the emotional readiness for change factor. 7KH&URQEDFK¶V alpha value increased from Į .34 to Į .80 when leaving the three items of cognitive

readiness for change out. Only two items will therefore assess cognitive readiness for change. Intentional readiness for change showed no inconsistencies within the factor analysis. The FURQEDFK¶VDOSKDYDOXHIRUWKLVVXE-construct was Į .86.

Implementation outcome

As mentioned before, different self-rating items will assess the implementation outcome for the quantitative part of this thesis. The EFA showed that items loaded on two different

(19)

19 After performing the above described data reduction, regression analyses were conducted in order to come to answers to the research questions. The results of those tests will be presented in chapter five of this thesis.

To explore the relations between the concepts of this study, different analyses were performed. First of al Pearson correlations were used to analyse the association of the variables with each other. Secondly, multiple regression analyses were performed to analyse the mediation effect (Baron & Kenny, 1986) and lastly, multiple hierarchical regression analyses were performed to explore the moderation.

3.7 Data analysis for the department level sub-study

For the department level results the focus lies on the experienced exposure to the

implementation interventions, the risk type functional uncertainty and the usage percentages of SURPASS.

Implementation interventions

An ANOVA analysis in combination with a post-hoc test was done to explore the differences in exposure to implementation interventions between the departments.

Functional uncertainty normally is a measure on implementation level. However, in this study, the mean scores for the risk type functional uncertainty was calculated for the departments to explore the differences between the departments.

Usage percentages

(20)

20

Individual level results

The data reduction did not result in new conceptual ideas for this study, which means that the two models presented earlier in chapter 2.5 are still under investigation. In this chapter, the results for the individual level sub-study will be presented.

4.1 Descriptive statistics and correlations

The number of respondents in this research was 165 out of 573 questionnaires that have been spread around the participating departments. The response rate is 28,8% (n = 165) of which 66 were male and 97 were female. The average age of the respondents was 39.6 years (SD= 11.2).In table 4 the categorical variables of the respondents are displayed, to look at

representativeness of the final set of respondents for the employees in the UMCG. In advance the division of respondents in female/male and in the different professions was known for the population. There is no indication of selective non-response based on these numbers. For the other measures, there was no information available for the sample or the population. Selective non-response may have occurred, however, it seems that all characteristics are well

represented.

Table 4. Respondents characteristics summary______________________________________ Respondents Sampled Population Percentage Percentage Percentage

(21)

(22)

22 Table 5. Descriptive statistics of the research variables

Means (M), standard deviations (SD) and correlation between the research variables

(23)

4.2 Moderator: Team climate

The results of the analyses for the expected moderation of team climate on the relation between unfreeze implementation interventions and the different constructs of readiness for change will be presented here. To analyze the impact of a moderator, a multiple hierarchical regression needs to be performed. Since the dependent variable in this analysis was divided into three dimensions, the test needs to be performed three times. The results of these analyses are shown in the tables 6, 7 and 8.

Table 6. Multiple hierarchical regression analysis (1)

Emotional readiness for change

Step 1 Step 2 Step 3 Age Gender Unfreeze TCI Interaction R2 ǻ52 .147 -.007 .022 .526* .112 .339* .317 .525* .109 -.008 .339 .000

Note: Step 1 = multiple regression analysis with control variables, Step 2 = multiple regression analysis with control variables and unfreeze elements and team climate, Step 3 = multiple hierarchical regression analysis with interaction term.

*P<.001

Table 7. Multiple hierarchical regression analysis (2)

Cognitive readiness for change

Step 1 Step 2 Step 3 Age Gender Unfreeze TCI Interaction R2 ǻ52 -.026 .056 .005 .268* .043 .083* .078 .268* .043 .001 .083 .000

Note: Step 1 = multiple regression analysis with control variables, Step 2 = multiple regression analysis with control variables and unfreeze elements and team climate, Step 3 = multiple hierarchical regression analysis with interaction term.

(24)

24 Table 8. Multiple hierarchical regression analysis (3)

Intentional readiness for change

Step 1 Step 2 Step 3 Age Gender Unfreeze TCI Interaction R2 ǻ52 .206** .007 .042 .124 .134 .084** .043 .152 .235* .268* .142* .058

Note: Step 1 = multiple regression analysis with control variables, Step 2 = multiple regression analysis with control variables and unfreeze interventions and team climate, Step 3 = multiple hierarchical regression analysis with interaction term.

*P<.01 ** P < .05

The outcomes of the analyses show that the moderating effect of team climate only received support for the relation between unfreeze implementation interventions and intentional readiness for change in the expected direction. The amount of variance explained was almost 6%, whereas the analyses for both cognitive and emotional readiness for change explained 0% variance.

4.3 Mediation: unfreezing ± readiness for change ± implementation outcome

The mediation analyses were done following the four steps of Baron and Kenny (1986). For this study, three mediation analyses need to be performed. The results of the analyses are displayed in the tables 9, 10 and 11 and figures 3, 4 and 5.

Table 9. Multiple regression analysis with mediator emotional readiness for change Adoption of change

Step 1 Step 2 Step 3 Age

Gender Unfreeze

Emotional readiness for change R2 ǻ52 .138 -.036 .023 .265* .091* .068 .018 .436* .217* .127

Note: Step 1 = multiple regression analysis with control variables, Step 2 = multiple regression analysis with control variables and unfreeze interventions, Step 3 = multiple regression analysis with emotional readiness for change as a mediator.

*P < .01

(25)

25 (0.19, p < .01) (.436, p < .01)

1. (0.265, p < .01) 2. (0.018, n.s.)

Figure 3. Mediation analysis for emotional readiness for change.

The results of the mediation analysis show that there is a partial mediation for emotional readiness for change between µunfreeze¶ implementation interventions and individual adoption to change.

Table 10. Multiple regression analysis with mediator cognitive readiness for change Adoption of change

Gender Unfreeze

Cognitive readiness for change R2 ǻ52 .138 -.036 .023 .265* .091* .068 .115 .528* .346* .256

Note: Step 1 = multiple regression analysis with control variables, Step 2 = multiple regression analysis with control variables and unfreeze interventions, Step 3 = multiple regression analysis with cognitive readiness for change as a mediator.

*P < .01

(.276, p < .01)

(0.528, p < .01)

1. (0.265, p < .01) 2. (0.115, n.s.)

Figure 4. Mediation analysis for cognitive readiness for change

The results of the mediation analysis show that there is a partial mediation for cognitive Unfreeze

elements change Adoption of

(26)

26 M ediator - Intentional readiness for change Unfreeze

elements change Adoption of

readiness for change between µunfreeze¶ implementation interventions and individual adoption of change.

Table 11. Multiple regression with mediator intentional readiness for change Adoption of change

Gender Unfreeze

Intentional readiness for change R2 ǻ52 .138 -.036 .023 .265* .091* .068 .211* .311* .181* .090

Note: Step 1 = multiple regression analysis with control variables, Step 2 = multiple regression analysis with control variables and unfreeze interventions, Step 3 = multiple regression analysis with emotional readiness for change as a mediator.

*P < .01

(0.190, p < .05) (0.311, p < .01)

1. (0.265, p < .01) 2. (0.211, p < .01)

Figure 5. Mediation analysis for intentional readiness for change

The results show that there is no mediation for intentional readiness for change between µXQIUHH]H¶LPSOHPentation interventions and individual adoption of change. Although mediation was not supported, there remains a significant direct relation between intentional readiness for change and adoption of change and between unfreeze implementation

interventions and adoption of change.

(27)

27 Table 12. Multiple regression with mediator intentional readiness for change

Adoption of change

Gender Unfreeze

Intentional readiness for change

Emotional readiness for change

Cognitive readiness for change R2 ǻ52 .138 -.036 .023 .265* .091* .068 .054 .186 .096 .438* .385* .294

Note: Step 1 = multiple regression analysis with control variables, Step 2 = multiple regression analysis with control variables and unfreeze interventions, Step 3 = multiple regression analysis with intentional readiness for change, emotional readiness for change and cognitive readiness for change as mediators.

(28)

28

Department level results

The purpose of this chapter is exploring the influence of implementation interventions and functional uncertainty on the implementation outcome (usage percentages of SURPASS). The results of the department level analyses will be presented and explained in light of the

available qualitative data. Department 1 and 2 are currently best performing departments, department 3 and 4 perform average and departments 5 and 6 are less performing departments (Table 2 in chapter 3 shows this case study sampling logic).

5.1 Comparison of department level experience of implementation interventions To compare the departments on the experience of implementation interventions, a one-way ANOVA test was conducted. This test shows if the expected differences between the

departments based on their implementation outcome can be explained by their differences in the experience of the implementation interventions. A Post-hoc test shows which groups, in this case departments, significantly differ. In the tables 13 and 14 the results of the test are displayed. A division between the implementation interventions was made between µunfreeze¶ items and µmove/refreeze¶ items, based on the explorative factor analysis as described before. Table 13. ANOVA test results for Unfreeze elements

Department Sample

size

Mean 1

2

3

4

5

6

1

35

3.77

x .59*

2

9

3.18

x -.75*

3

55

3.5

x

4

11

3.5

x

5

8

3.9

x

6

10

3.3

x * = p < .05, n = 128

Scale: 1 = totally disagree, 5 = totally agree (on the presence of an implementation intervention)

Table 14. ANOVA test results for Move/Refreeze elements

Department Sample

size

Mean 1

2

3

4

5

6

1

35

3.52

x .62* .29* .87*

2

9

2.9

x

3

55

3.23

x .58*

4

11

3.4

x .77*

5

8

3.1

x

6

10

2.66

x * = p < .05, n = 128

(29)

29 Expected was that the departments currently performing well would have a significant higher mean score than the average and less performing departments and that the average performing departments would have higher mean scores than less performing departments. However, the results for the unfreeze implementation interventions show that the distinction between the departments is not exactly as expected. Especially department two and five, a well performing and less performing department, show interesting results. Department two shows low mean scores on the exposure to both unfreeze and move/refreeze implementation interventions and differs significantly from the other good performing department on both unfreeze and

move/refreeze. Whereas department five shows relatively high scores on the experience of the unfreeze implementation interventions while their current usage percentage is low.

Looking at significant differences between the departments, not all of the outcomes from the ANOVA and post-hoc analyses are significant and thus not fully generalizable, however, looking at only this case, there are some consistencies worthy to mention.

The means for the unfreeze elements (except for department 2 and 5) are distributed as expected. Higher means show a more positive experience of the implementation

interventions. The means for the move/refreeze elements are also distributed as expected (except for department 2). Neglecting this deviating department, it means that the best

performing department has a more positive experience about the implementation interventions than the average performing departments and the average performing departments on their turn have a more positive experience about the implementation interventions than the worse performing department. For this study it thus seems that the experienced exposure to

implementation interventions has a direct effect on the usage percentage of the quality instrument.

5.2 Assessment of functional uncertainty as a moderator

In this section we are exploring the influence of functional uncertainty. Table 15 shows the scores on the different items of functional uncertainty. Those will be related to the

experienced exposure to the implementation interventions.

Table 15. Risk (functional uncertainty) in participating departments

Fa Fb Fc Fd

Department 1 Medium Medium Low Low

Department 2 Medium Medium Medium Low

Department 3 Medium Medium Low Medium

Department 4 Medium High Medium Medium

Department 5 Medium Medium Medium Medium

Department 6 Medium Medium Medium Med-high

(30)

30 Department two and three show slightly higher risk than department one. Department two, however, shows also lower mean scores on exposure to the implementation interventions. Than the fact that they still perform well in terms of SURPASS usage is thus not explained by the exposure to implementation interventions and functional uncertainty and might be due to other factors. For department 3 the logic might count that their moderate exposure experience to implementation interventions and their medium scores on functional uncertainty explain their moderate usage percentage.

Department 4, 5 and 6 show almost similar scores for functional uncertainty (medium-high). The exposure to implementation interventions was consistent with their usage percentage (lowest exposure ± lowest usage percentage and vice versa). This might mean that the functional uncertainty they experience does have equal impact in the departments and their outcome differences might thus only be explained by their exposure to the implementation interventions.

From the implementation reports and informal conversations it appeared that several departments complained about the fact that their whishes for adjustments to the SURPASS checklist were not encountered and not taken seriously. Different key people in the

implementation also found that the information stream concerning SURPASS was inadequate and that SURPASS was hard to tailor into their daily work. The available documents about the SURPASS implementation show that little time was given to adjust the processes within departments. In appendix D an overview of the planned and achieved implementation interventions is given (source: Implementatieplan 2.0). Little planned interventions were actually done as planned which may explain the relatively high risk type functional

uncertainty. )XUWKHUPRUHQRDWWHQWLRQZDVGUDZQWRDFHUWDLQµFKDQJHVW\OH¶DQGNH\SHRSOH in the implementation were able to follow their own ideas, which might have caused the differences between departments. Koopman & Pool (1991) suggest a decision making model IRUFKDQJHVZKHUHLQDSDWKVKRXOGEHIROORZHGWKDWILWVWKHµFKDQJH¶KHUHquality instrument implementation) that is going to be implemented. From own experience and conversations with implementation coordinators, it appears that a path was followed where little

participation was possible and decisions were made top-down, even though participation from bottom-up could have highlighted problems and could have provided crucial information about workflows in departments and coordination needs between departments. The fact that the departments score high on the risk type functional uncertainty means that participation was necessary to achieve more successful implementation outcomes (higher usage

percentages). The model of Koopman & Pool (1991) is displayed in appendix E. 5.3 Outcome

The exposure to implementation interventions shows some consistencies, except for department 2. The means are distributed as expected and even though not all differences between the departments are significant, for this study they can be explained. The moderating effect as proposed in the model for the department level sub-study is not fully supported. However, functional uncertainty does certainly plays a role in the success of an

(31)

31

Discussion

Previous research has already indicated that contingencies play a role in organizational change projects, such as quality instrument implementations. However, little research is available on which contingencies that are of influence. This research explored the influence of two contingencies on the outcome of a quality instrument implementation in a healthcare setting. Furthermore, it was explored how implementation interventions and readiness for change play a role in reaching the implementation goal: adoption of change. The study was divided in two sub-studies, one on individual level and one on departmental level. Several relationships from the conceptual model for the individual level sub-study were found to be significant. The results for the departmental level sub-study showed consistencies, but these were not significant and, therefore, it is for now not possible to generalize the results. However, functional uncertainty and the exposure to implementation interventions seem to play a predictive role for implementation outcome.

The results and their implications will be discussed more thoroughly in this chapter. Moderator team climate

Team climate was expected to moderate the effect of WKHµXQIUHH]H¶implementation interventions on the different readiness for change dimensions. This moderation was only found for intentional readiness for change, and not for the cognitive and emotional readiness for change.

A possible explanation for this finding may be found in the different dimensions of readiness for change all focusing on specific aspects of readiness for change. Emotional readiness and cognitive readiness for change UHSUHVHQWDQLQGLYLGXDO¶Vdeeply embedded feelings and perceptions about change in general (Bouckenhooge et al., 2009). The construct intentional readiness for change represents the action that an individual is willing to put in a particular change. The intention to act might be easier influenced by a team¶Vclimate than VRPHRQH¶VGHHSHUKHOGEHOLHIVDERXWFKDQJHs.

In addition to this explanation, the underlying reason for this unexpected finding may be found in the social identity and influence/persuasion theories. Wood (2000) describes the factors that may change the behaviour of individuals in a particular situation. The social identity theory (Tajfel, 1982) shows how group influences work. When people categorize themselves as in-group members, the in-group becomes a reference for social comparison and people adopt the attitudes and beliefs of the in-group as if they are their own. This means that either positive or negative attitudes of team members towards in this case the quality

instrument, influences other team members that feel like an in-group team member and maybe even act like team members.

(32)

32 Mediator readiness for change

Partial mediation was found for both emotional and cognitive readiness for change. Regarding intentional readiness for change the mediation did not occur. :KHQWKHWKUHHµUHDGLQHVVIRU FKDQJH¶GLPHQVLRQVZHUHanalysed together, the only significant effect that remained was mediation for cognitive readiness for change.

A possible explanation for this finding may be found looking at the nature of the quality instrument under investigation (SURPASS: a checklist throughout the surgical trajectory to reduce adverse events and increase patient safety). According to William Labov (2010) cognition its most general sense, denotes any form of knowing. ³The most relevant definition RIFRJQLWLRQLV³WKHDFWLRQRUIDFXOW\RINQRZLQJWDNHQLQLWVZLGHVWVHQVHLQFOXGLQJVHQVDWLRQ SHUFHSWLRQFRQFHSWLRQHWFDVGLVWLQJXLVKHGIURPIHHOLQJDQGYROLWLRQ´ Employees

confronted with the quality instrument under investigation, probably know that it is indeed best to use it for the patient safety and they seemed to have used their cognition to change their behavior. A problem that can arise in this process is cognitive dissonance. Cognitive dissonance is the situation when people sense an inconsistency either between two or more attitudes or between their attitudes and behavior; that is, they will feel frustrated and uncomfortable ± sometimes extremely so ± with the situation (Burnes & James, 1994). Therefore, individuals will seek a stable state where there is minimum dissonance, moreover, the level and type of involvement should be geared to the level of dissonance that any

proposed changes may give rise to (Burnes & James, 1994) to increase implementation success.

For results of the mediation analyses it should be taken into account that the measure for adoption to change was a self-rating one and was quite subjective. Future research should determine whether the mediation still holds when objective measures are used. Furthermore, as described above, some problems arose with the CFA, which could have affected the outcome for the mediation analyses as well.

Implementation interventions and functional uncertainty

In the model for the departmental level sub-study, functional uncertainty was conceptualized as a moderator between implementation interventions and the adoption of change. Functional uncertainty is a concept from the RABSODY instrument (van Offenbeek & Koopman, 1996). This instrument has been adapted to diagnose risk factors in implementation of technical innovations within the healthcare sector (Jenner et al., 2008). The risk factors should actually be assessed in advance of the implementation to choose fitting change interventions. In this research, the assessment was done during the implementation and not used to answer the risks by a matching implementation strategy. The six participating departments all show relatively high risk scores, which may mean that the need for exploring functional uncertainty in this implementation existed and that changes in the implementation interventions could have easily increased the innovation-system fit.

There are consistent findings looking at the distribution of the means of

(33)

33 functional uncertainty than the worst performing departments. However, not all the findings were found to be significant, which has its implications for the generalizability of the results.

The fact that department two, which scores high on adoption of change, does not score high on the exposure to implementation interventions and moreover has a medium amount of functional uncertainty needs additional explanation. A possible explanation may be found in a deviating leadership style of the department leader or key figures compared to other

GHSDUWPHQWV¶OHDGHUV$GLUHFWLYHRUHYHQFRHUFLYHVW\OH'XQSK\DQG6WDFHZKHUHLQ QRWDFWLQJDFFRUGLQJWRWKHµUXOHV¶is simply not a possibility, seems to have been present. The Dunphy and Stace matrix (1993) shows different styles of change management that fit with different change projects, depending on the context. Future research is needed to determine the influence of leadership style as a possible third contingency in implementations of quality instruments in healthcare settings.

Limitations and strengths

A number of limitations should be mentioned about this research. First of all, the selection of contingency factors under investigation was not exhaustive; future research should determine if other contingency factors might also be of influence on the success of implementations in healthcare settings. For example the other risks that are in the RABSODY instrument (van Offenbeek & Koopman, 1996). In this research functional uncertainty seemed most relevant, however, there are five risk types to be diagnosed in advance of an implementation according to the instrument. As described above, also the influence of leadership would be an interesting contingency to research.

In the methodological part of this research, some limitations are also present. The response percentage of the research was lower than expected: for the individual level sub-study 28,8% and for the department level sub-sub-study 22,3%, which had its consequences. For the sub-study on department level it was at times hard to interpret the data, since response percentages of several departments were below 10% (8 employees). This may mean that only the most positive or most negative employees participated in the research and that the data are not representative. Interviews with several key figures of departments would have been interesting to find underlying reasons of the problems or negative feelings towards the implementation. With the individual sub-study of this research some problems arose as well. The confirmatory factor analysis showed that seven questionnaire items could not be included in the analysis, despite the fact that the questionnaire was largely based on validated

(34)

34 departmental sub-study of this research can be generalized is questionable. Future research in other settings may provide an answer to this question.

Besides the limitations of this research, strengths are also worthwhile to mention. The combination between qualitative and quantitative data provided insights and explanations for this case; Eisenhardt (1989) underlines the synergy of both quantitative and qualitative

research. Furthermore, the two sub-studies on different levels provided the opportunity to find logic in both individual and departmental results. The study on the two different levels made it possible to exclude part of the social desirability bias from the self-rating measure for the implementation outcome, since an objective measure (usage percentages of the quality instrument) was used for department-level comparisons.

Conclusion and managerial implications

The aim of this study was to provide information about the influence of two contingencies on quality instrument implementations in healthcare settings and to explore a possible mediation of readiness for change between implementation interventions and the adoption of change.

To conclude, the implementation interventions were found to have significant

influence on the adoption of change. This relation was mediated only by cognitive readiness for change, which seems plausible looking at the nature of the quality instrument under investigation. Furthermore, this study found the moderating effect of team climate on the UHODWLRQEHWZHHQµXQIUHH]H¶LPSOHPHQWDWLRQLQWHUYHQWLRQVDQGintentional readiness for change. This effect shows the influence team members have on other members of the same team. The assessment of functional uncertainty showed some consistencies comparing the mean scores of the participating departments. Moreover, the risk type functional uncertainty was quite high for all departments, which can be explained for this particular setting, since no assessment of the risk was done in advance of the implementation and no action was taken to reduce the amount of risk perceived. Future research is needed to determine the exact

influence of this risk on the implementation outcomes in general.

(35)

35 recommendation is that implementation interventions should be tailored to the type of quality instrument. In this particular case, cognitive readiness for change was found to have high impact on the relation between implementation interventions and the adoption of change. Interventions influencing the cognitive readiness for change seem to lead to a higher adoption of change

(36)

36

References

Anderson, N.R. & West, M.A. 1998. Measuring climate for work group innovation: development and validation of the team climate inventory. Journal of

Organizational Behavior. 19:235-258.

Armenakis, A.A., Harris, S.G. & Mossholder, K.W. 1993. Creating readiness for organizational change. H uman Relations, 46:6, 681-703

Basit, T.N. 2003. Manual or electronic? The role of coding in qualitative data analysis. Educational Research, 45(2), 143-154

Bouckenhooghe, D., Devos, G.,van den Broeck, H. 2009. Organizational Change

Questionnaire±Climate of Change, Processes, and Readiness: Development of a New Instrument. The Journal of Psychology, 143:6, 559±599

Bower, P., Campbell, S., Bojke, C. & Sibbald, B. 2003. Team structure, team climate and the quality of care in primary care: an observational study. Quality Safety Health Care. 12:273-279

Buchanan, D., Fitzgerals, L., Ketley, D., Gollop, R., Jones, J.L., Saint Lamont, S., Neath, A. & Whitby, E. 2005. No going back: A review of the literature on sustaining organizational change. International Journal of Management Reviews. 7:3, 189-205

Burnes, B., 2004(a). Emergent change and planned change ± competitors or allies? The case of XYZ construction, International Journal of Operations & Product

Management, 24: 886-902.

Burnes, B., 2004(b). Kurt Lewin and the Planned Approach to Change: A Re-appraisal, Journal of Management Studies 41:6. 0022-2380

Burnes, B. 2009. Managing Change (5th ed.). Pearson Education Limited.

Burnes, B. & James, H., 1995. Culture, cognitive dissonance and the management of change. International Journal of Operations & Production Management, 15: 8, 14 ± 33. Cummings, T.G. & Worley, C.G. 2008. Organizational Development & Change (9th ed.).

Mason: Cengage Learning

(37)

37 Dunphy, D. & Stace, D. 1993. The strategic Management of Corporate Change. H uman

Relations, 46:8, 905-920.

Edmondson, A. 1999. Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44: 350-383.

Edwards, J.R. and Bagozzi, R.P. 2000. On the Nature and Direction of Relationships Between Constructs and Measures. Psychological Methods. 5:2, 155-174. Eisenhardt, K.M. 1989. Building Theories from Case Study Research. Acadamy of

Management Review. 4:532-550

Fishbein, M. & Ajzen, I. 1975. Belief, Attitude, Intention and Behaviour: An Introduction to Theory and Research. Reading MA AddisonWesley.

Glisson, C. and James, L.R. 2002. The cross-level effects of culture and climate in human service teams. Journal of Organizational Behavior, 23: 767±794

Greenhalgh, T., Robert, G., Macfarlane, F., Bate, P. & Kyriakidou, O. 2004. Diffusion of innovations in service organizations: systematic review and recommendations. The Milbank Quarterly. 82,4,581-629.

Grol & Wensing 2005. Implementatie: effectieve verbetering van de patiëntenzorg. Elsevier gezondheidszorg, 2006.

Holt, D. T., Armenakis,A., Feild, H. S.,& Harris, S. G. 2007. Readiness for organizational change: The systematic development of a scale. Journal of Applied Behavioral Science, 43, 232±255.

Hulscher, M.E.J.L., Laurant, M.G.H. & Grol, R.P.T.M. 2002. Process evaluation on quality improvement interventions. Quality improvement research. 40-46. Jenner, J.A., Stant, A.D., van Offenbeek, M.A.G., ten Vergert, E. & Wiersma, D. 2008.

Implementatie van de nieuwe kosteneffectieve behandeling HIT voor patiënten met therapieresistente gehoorshallucinaties. E indverslag ZonMw.

Jones R.A., Jimmieson, N.L. & Griffiths, A. 2005. The Impact of Organizational Culture and Reshaping Capabilities on Change Implementation Success: The Mediating Role of Readiness for Change. Journal of Management Studies, 42(2): 363-383

An overall desired goal does not make the change -